* [PATCH net-next 08/12] be2net: define BE_MAX_MTU
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Kalesh AP <kalesh.purayil@emulex.com>
This patch defines a new macro BE_MAX_MTU to make the code in be_change_mtu()
more readable.
Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be.h | 2 ++
drivers/net/ethernet/emulex/benet/be_main.c | 14 +++++++-------
2 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 069bdda..5ccfce7 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -86,6 +86,8 @@ static inline char *nic_name(struct pci_dev *pdev)
#define BE_MAX_JUMBO_FRAME_SIZE 9018
#define BE_MIN_MTU 256
+#define BE_MAX_MTU (BE_MAX_JUMBO_FRAME_SIZE - \
+ (ETH_HLEN + ETH_FCS_LEN))
#define BE_NUM_VLANS_SUPPORTED 64
#define BE_MAX_EQD 128u
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index cdad8bc..04a8092 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -1073,15 +1073,15 @@ static netdev_tx_t be_xmit(struct sk_buff *skb, struct net_device *netdev)
static int be_change_mtu(struct net_device *netdev, int new_mtu)
{
struct be_adapter *adapter = netdev_priv(netdev);
- if (new_mtu < BE_MIN_MTU ||
- new_mtu > (BE_MAX_JUMBO_FRAME_SIZE - (ETH_HLEN + ETH_FCS_LEN))) {
- dev_info(&adapter->pdev->dev,
- "MTU must be between %d and %d bytes\n",
- BE_MIN_MTU,
- (BE_MAX_JUMBO_FRAME_SIZE - (ETH_HLEN + ETH_FCS_LEN)));
+ struct device *dev = &adapter->pdev->dev;
+
+ if (new_mtu < BE_MIN_MTU || new_mtu > BE_MAX_MTU) {
+ dev_info(dev, "MTU must be between %d and %d bytes\n",
+ BE_MIN_MTU, BE_MAX_MTU);
return -EINVAL;
}
- dev_info(&adapter->pdev->dev, "MTU changed from %d to %d bytes\n",
+
+ dev_info(dev, "MTU changed from %d to %d bytes\n",
netdev->mtu, new_mtu);
netdev->mtu = new_mtu;
return 0;
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 06/12] be2net: fix log messages in lancer FW download path
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Kalesh AP <kalesh.purayil@emulex.com>
Log messages in the Lancer FW download path have issues such as:
- a single message spanning multiple lines
- the success message is logged even in failure cases
- status codes are already logged in the FW cmd routines
This patch fixes these issues.
Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be_main.c | 30 +++++++++-----------------
1 files changed, 11 insertions(+), 19 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index c95998f..7e89e5b 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -4076,6 +4076,7 @@ static int lancer_fw_download(struct be_adapter *adapter,
{
#define LANCER_FW_DOWNLOAD_CHUNK (32 * 1024)
#define LANCER_FW_DOWNLOAD_LOCATION "/prg"
+ struct device *dev = &adapter->pdev->dev;
struct be_dma_mem flash_cmd;
const u8 *data_ptr = NULL;
u8 *dest_image_ptr = NULL;
@@ -4088,16 +4089,14 @@ static int lancer_fw_download(struct be_adapter *adapter,
u8 change_status;
if (!IS_ALIGNED(fw->size, sizeof(u32))) {
- dev_err(&adapter->pdev->dev,
- "FW Image not properly aligned. "
- "Length must be 4 byte aligned.\n");
+ dev_err(dev, "FW image size should be multiple of 4\n");
status = -EINVAL;
goto lancer_fw_exit;
}
flash_cmd.size = sizeof(struct lancer_cmd_req_write_object)
+ LANCER_FW_DOWNLOAD_CHUNK;
- flash_cmd.va = dma_alloc_coherent(&adapter->pdev->dev, flash_cmd.size,
+ flash_cmd.va = dma_alloc_coherent(dev, flash_cmd.size,
&flash_cmd.dma, GFP_KERNEL);
if (!flash_cmd.va) {
status = -ENOMEM;
@@ -4137,33 +4136,26 @@ static int lancer_fw_download(struct be_adapter *adapter,
&add_status);
}
- dma_free_coherent(&adapter->pdev->dev, flash_cmd.size, flash_cmd.va,
- flash_cmd.dma);
+ dma_free_coherent(dev, flash_cmd.size, flash_cmd.va, flash_cmd.dma);
if (status) {
- dev_err(&adapter->pdev->dev,
- "Firmware load error. "
- "Status code: 0x%x Additional Status: 0x%x\n",
- status, add_status);
+ dev_err(dev, "Firmware load error\n");
goto lancer_fw_exit;
}
+ dev_info(dev, "Firmware flashed successfully\n");
+
if (change_status == LANCER_FW_RESET_NEEDED) {
- dev_info(&adapter->pdev->dev,
- "Resetting adapter to activate new FW\n");
+ dev_info(dev, "Resetting adapter to activate new FW\n");
status = lancer_physdev_ctrl(adapter,
PHYSDEV_CONTROL_FW_RESET_MASK);
if (status) {
- dev_err(&adapter->pdev->dev,
- "Adapter busy for FW reset.\n"
- "New FW will not be active.\n");
+ dev_err(dev, "Adapter busy, could not reset FW\n");
+ dev_err(dev, "Reboot server to activate new FW\n");
goto lancer_fw_exit;
}
} else if (change_status != LANCER_NO_RESET_NEEDED) {
- dev_err(&adapter->pdev->dev,
- "System reboot required for new FW to be active\n");
+ dev_info(dev, "Reboot server to activate new FW\n");
}
-
- dev_info(&adapter->pdev->dev, "Firmware flashed successfully\n");
lancer_fw_exit:
return status;
}
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 09/12] be2net: make be_cmd_get_regs() return a status
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Vasundhara Volam <vasundhara.volam@emulex.com>
There are a few failure cases in be_cmd_get_regs() that ideally must return
an error value. This style is used across all the routines in be_cmds.c with
this routine being an exception. This patch fixes this.
Signed-off-by: Vasundhara Volam <vasundhara.volam@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be_cmds.c | 10 +++++-----
drivers/net/ethernet/emulex/benet/be_cmds.h | 2 +-
2 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index ca9ec18..5be100d 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c
@@ -1681,17 +1681,17 @@ err:
return status;
}
-void be_cmd_get_regs(struct be_adapter *adapter, u32 buf_len, void *buf)
+int be_cmd_get_regs(struct be_adapter *adapter, u32 buf_len, void *buf)
{
struct be_dma_mem get_fat_cmd;
struct be_mcc_wrb *wrb;
struct be_cmd_req_get_fat *req;
u32 offset = 0, total_size, buf_size,
log_offset = sizeof(u32), payload_len;
- int status;
+ int status = 0;
if (buf_len == 0)
- return;
+ return -EIO;
total_size = buf_len;
@@ -1700,10 +1700,9 @@ void be_cmd_get_regs(struct be_adapter *adapter, u32 buf_len, void *buf)
get_fat_cmd.size,
&get_fat_cmd.dma);
if (!get_fat_cmd.va) {
- status = -ENOMEM;
dev_err(&adapter->pdev->dev,
"Memory allocation failure while retrieving FAT data\n");
- return;
+ return -ENOMEM;
}
spin_lock_bh(&adapter->mcc_lock);
@@ -1746,6 +1745,7 @@ err:
pci_free_consistent(adapter->pdev, get_fat_cmd.size,
get_fat_cmd.va, get_fat_cmd.dma);
spin_unlock_bh(&adapter->mcc_lock);
+ return status;
}
/* Uses synchronous mcc */
diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.h b/drivers/net/ethernet/emulex/benet/be_cmds.h
index 5284b82..0e11868 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.h
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.h
@@ -2101,7 +2101,7 @@ int be_cmd_get_die_temperature(struct be_adapter *adapter);
int be_cmd_get_cntl_attributes(struct be_adapter *adapter);
int be_cmd_req_native_mode(struct be_adapter *adapter);
int be_cmd_get_reg_len(struct be_adapter *adapter, u32 *log_size);
-void be_cmd_get_regs(struct be_adapter *adapter, u32 buf_len, void *buf);
+int be_cmd_get_regs(struct be_adapter *adapter, u32 buf_len, void *buf);
int be_cmd_get_fn_privileges(struct be_adapter *adapter, u32 *privilege,
u32 domain);
int be_cmd_set_fn_privileges(struct be_adapter *adapter, u32 privileges,
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 04/12] be2net: Add TX completion error statistics in ethtool
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Kalesh AP <kalesh.purayil@emulex.com>
HW reports TX completion errors in TX completion. This patch adds these
counters to ethtool statistics.
Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be.h | 7 ++++
drivers/net/ethernet/emulex/benet/be_ethtool.c | 28 +++++++++++++++
drivers/net/ethernet/emulex/benet/be_hw.h | 12 ++++++
drivers/net/ethernet/emulex/benet/be_main.c | 45 ++++++++++++++++++++++++
4 files changed, 92 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index d9cd885..d491ac6 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -248,6 +248,13 @@ struct be_tx_stats {
ulong tx_jiffies;
u32 tx_stops;
u32 tx_drv_drops; /* pkts dropped by driver */
+ /* the error counters are described in be_ethtool.c */
+ u32 tx_hdr_parse_err;
+ u32 tx_dma_err;
+ u32 tx_tso_err;
+ u32 tx_spoof_check_err;
+ u32 tx_qinq_err;
+ u32 tx_internal_parity_err;
struct u64_stats_sync sync;
struct u64_stats_sync sync_compl;
};
diff --git a/drivers/net/ethernet/emulex/benet/be_ethtool.c b/drivers/net/ethernet/emulex/benet/be_ethtool.c
index d8d7a4a..3a7ade4 100644
--- a/drivers/net/ethernet/emulex/benet/be_ethtool.c
+++ b/drivers/net/ethernet/emulex/benet/be_ethtool.c
@@ -157,6 +157,34 @@ static const struct be_ethtool_stat et_rx_stats[] = {
*/
static const struct be_ethtool_stat et_tx_stats[] = {
{DRVSTAT_TX_INFO(tx_compl)}, /* If moving this member see above note */
+ /* This counter is incremented when the HW encounters an error while
+ * parsing the packet header of an outgoing TX request. This counter is
+ * applicable only for BE2, BE3 and Skyhawk based adapters.
+ */
+ {DRVSTAT_TX_INFO(tx_hdr_parse_err)},
+ /* This counter is incremented when an error occurs in the DMA
+ * operation associated with the TX request from the host to the device.
+ */
+ {DRVSTAT_TX_INFO(tx_dma_err)},
+ /* This counter is incremented when MAC or VLAN spoof checking is
+ * enabled on the interface and the TX request fails the spoof check
+ * in HW.
+ */
+ {DRVSTAT_TX_INFO(tx_spoof_check_err)},
+ /* This counter is incremented when the HW encounters an error while
+ * performing TSO offload. This counter is applicable only for Lancer
+ * adapters.
+ */
+ {DRVSTAT_TX_INFO(tx_tso_err)},
+ /* This counter is incremented when the HW detects Q-in-Q style VLAN
+ * tagging in a packet and such tagging is not expected on the outgoing
+ * interface. This counter is applicable only for Lancer adapters.
+ */
+ {DRVSTAT_TX_INFO(tx_qinq_err)},
+ /* This counter is incremented when the HW detects parity errors in the
+ * packet data. This counter is applicable only for Lancer adapters.
+ */
+ {DRVSTAT_TX_INFO(tx_internal_parity_err)},
{DRVSTAT_TX_INFO(tx_bytes)},
{DRVSTAT_TX_INFO(tx_pkts)},
/* Number of skbs queued for trasmission by the driver */
diff --git a/drivers/net/ethernet/emulex/benet/be_hw.h b/drivers/net/ethernet/emulex/benet/be_hw.h
index 8840c64..295ee08 100644
--- a/drivers/net/ethernet/emulex/benet/be_hw.h
+++ b/drivers/net/ethernet/emulex/benet/be_hw.h
@@ -315,6 +315,18 @@ struct be_eth_hdr_wrb {
u32 dw[4];
};
+/********* Tx Compl Status Encoding *********/
+#define BE_TX_COMP_HDR_PARSE_ERR 0x2
+#define BE_TX_COMP_NDMA_ERR 0x3
+#define BE_TX_COMP_ACL_ERR 0x5
+
+#define LANCER_TX_COMP_LSO_ERR 0x1
+#define LANCER_TX_COMP_HSW_DROP_MAC_ERR 0x3
+#define LANCER_TX_COMP_HSW_DROP_VLAN_ERR 0x5
+#define LANCER_TX_COMP_QINQ_ERR 0x7
+#define LANCER_TX_COMP_PARITY_ERR 0xb
+#define LANCER_TX_COMP_DMA_ERR 0xd
+
/* TX Compl Queue Descriptor */
/* Pseudo amap definition for eth_tx_compl in which each bit of the
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 810ce5c..e9658f7 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -2421,11 +2421,49 @@ loop_continue:
return work_done;
}
+static inline void be_update_tx_err(struct be_tx_obj *txo, u32 status)
+{
+ switch (status) {
+ case BE_TX_COMP_HDR_PARSE_ERR:
+ tx_stats(txo)->tx_hdr_parse_err++;
+ break;
+ case BE_TX_COMP_NDMA_ERR:
+ tx_stats(txo)->tx_dma_err++;
+ break;
+ case BE_TX_COMP_ACL_ERR:
+ tx_stats(txo)->tx_spoof_check_err++;
+ break;
+ }
+}
+
+static inline void lancer_update_tx_err(struct be_tx_obj *txo, u32 status)
+{
+ switch (status) {
+ case LANCER_TX_COMP_LSO_ERR:
+ tx_stats(txo)->tx_tso_err++;
+ break;
+ case LANCER_TX_COMP_HSW_DROP_MAC_ERR:
+ case LANCER_TX_COMP_HSW_DROP_VLAN_ERR:
+ tx_stats(txo)->tx_spoof_check_err++;
+ break;
+ case LANCER_TX_COMP_QINQ_ERR:
+ tx_stats(txo)->tx_qinq_err++;
+ break;
+ case LANCER_TX_COMP_PARITY_ERR:
+ tx_stats(txo)->tx_internal_parity_err++;
+ break;
+ case LANCER_TX_COMP_DMA_ERR:
+ tx_stats(txo)->tx_dma_err++;
+ break;
+ }
+}
+
static bool be_process_tx(struct be_adapter *adapter, struct be_tx_obj *txo,
int budget, int idx)
{
struct be_eth_tx_compl *txcp;
int num_wrbs = 0, work_done;
+ u32 compl_status;
for (work_done = 0; work_done < budget; work_done++) {
txcp = be_tx_compl_get(&txo->cq);
@@ -2434,6 +2472,13 @@ static bool be_process_tx(struct be_adapter *adapter, struct be_tx_obj *txo,
num_wrbs += be_tx_compl_process(adapter, txo,
GET_TX_COMPL_BITS(wrb_index,
txcp));
+ compl_status = GET_TX_COMPL_BITS(status, txcp);
+ if (compl_status) {
+ if (lancer_chip(adapter))
+ lancer_update_tx_err(txo, compl_status);
+ else
+ be_update_tx_err(txo, compl_status);
+ }
}
if (work_done) {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 01/12] be2net: add a few log messages
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
This patch adds the following log messages to help debugging
failure cases:
1) log FW version number: this is useful when driver initialization
fails and the FW version number cannot be queried via ethtool
2) per function resource limits for BEx chips: these values are
currently being printed only for Skyhawk and Lancer
3) PCI BAR mapping failure
4) function_mode/caps queried from FW: this helps catch any FW bugs
that could advertise wrong capabilities to the driver
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be_cmds.c | 4 ++++
drivers/net/ethernet/emulex/benet/be_main.c | 20 ++++++++++++--------
2 files changed, 16 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index 4370ec1..ca9ec18 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c
@@ -1771,6 +1771,7 @@ int be_cmd_get_fw_ver(struct be_adapter *adapter)
status = be_mcc_notify_wait(adapter);
if (!status) {
struct be_cmd_resp_get_fw_version *resp = embedded_payload(wrb);
+
strcpy(adapter->fw_ver, resp->firmware_version_string);
strcpy(adapter->fw_on_flash, resp->fw_on_flash_version_string);
}
@@ -2018,6 +2019,9 @@ int be_cmd_query_fw_cfg(struct be_adapter *adapter)
adapter->function_mode = le32_to_cpu(resp->function_mode);
adapter->function_caps = le32_to_cpu(resp->function_caps);
adapter->asic_rev = le32_to_cpu(resp->asic_revision) & 0xFF;
+ dev_info(&adapter->pdev->dev,
+ "FW config: function_mode=0x%x, function_caps=0x%x\n",
+ adapter->function_mode, adapter->function_caps);
}
mutex_unlock(&adapter->mbox_lock);
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 93ff8ef..3d33049 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -3413,16 +3413,16 @@ static int be_get_resources(struct be_adapter *adapter)
if (be_roce_supported(adapter))
res.max_evt_qs /= 2;
adapter->res = res;
-
- dev_info(dev, "Max: txqs %d, rxqs %d, rss %d, eqs %d, vfs %d\n",
- be_max_txqs(adapter), be_max_rxqs(adapter),
- be_max_rss(adapter), be_max_eqs(adapter),
- be_max_vfs(adapter));
- dev_info(dev, "Max: uc-macs %d, mc-macs %d, vlans %d\n",
- be_max_uc(adapter), be_max_mc(adapter),
- be_max_vlans(adapter));
}
+ dev_info(dev, "Max: txqs %d, rxqs %d, rss %d, eqs %d, vfs %d\n",
+ be_max_txqs(adapter), be_max_rxqs(adapter),
+ be_max_rss(adapter), be_max_eqs(adapter),
+ be_max_vfs(adapter));
+ dev_info(dev, "Max: uc-macs %d, mc-macs %d, vlans %d\n",
+ be_max_uc(adapter), be_max_mc(adapter),
+ be_max_vlans(adapter));
+
return 0;
}
@@ -3633,6 +3633,7 @@ static int be_setup(struct be_adapter *adapter)
goto err;
be_cmd_get_fw_ver(adapter);
+ dev_info(dev, "FW version is %s\n", adapter->fw_ver);
if (BE2_chip(adapter) && fw_major_num(adapter->fw_ver) < 4) {
dev_err(dev, "Firmware on card is old(%s), IRQs may not work.",
@@ -4506,6 +4507,7 @@ static int be_map_pci_bars(struct be_adapter *adapter)
return 0;
pci_map_err:
+ dev_err(&adapter->pdev->dev, "Error in mapping PCI BARs\n");
be_unmap_pci_bars(adapter);
return -ENOMEM;
}
@@ -4822,6 +4824,8 @@ static int be_probe(struct pci_dev *pdev, const struct pci_device_id *pdev_id)
struct net_device *netdev;
char port_name;
+ dev_info(&pdev->dev, "%s version is %s\n", DRV_NAME, DRV_VER);
+
status = pci_enable_device(pdev);
if (status)
goto do_none;
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 12/12] be2net: query max_tx_qs for BE3 super-nic profile from FW
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Suresh Reddy <Suresh.Reddy@emulex.com>
In the BE3 super-nic profile, the max_tx_qs value can vary for each function.
So the driver needs to query this value from FW instead of using the
pre-defined constant BE3_MAX_TX_QS.
Signed-off-by: Suresh Reddy <Suresh.Reddy@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be_main.c | 13 +++++++++++--
1 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index f3235d1..a30156b 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -3324,10 +3324,19 @@ static void BEx_get_resources(struct be_adapter *adapter,
*/
if (BE2_chip(adapter) || use_sriov || (adapter->port_num > 1) ||
!be_physfn(adapter) || (be_is_mc(adapter) &&
- !(adapter->function_caps & BE_FUNCTION_CAPS_RSS)))
+ !(adapter->function_caps & BE_FUNCTION_CAPS_RSS))) {
res->max_tx_qs = 1;
- else
+ } else if (adapter->function_caps & BE_FUNCTION_CAPS_SUPER_NIC) {
+ struct be_resources super_nic_res = {0};
+ /* On a SuperNIC profile, the driver needs to use the
+ * GET_PROFILE_CONFIG cmd to query the per-function TXQ limits
+ */
+ be_cmd_get_profile_config(adapter, &super_nic_res, 0);
+ /* Some old versions of BE3 FW don't report max_tx_qs value */
+ res->max_tx_qs = super_nic_res.max_tx_qs ? : BE3_MAX_TX_QS;
+ } else {
res->max_tx_qs = BE3_MAX_TX_QS;
+ }
if ((adapter->function_caps & BE_FUNCTION_CAPS_RSS) &&
!use_sriov && be_physfn(adapter))
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 10/12] be2net: get rid of TX budget
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
Enforcing a budget on the TX completion processing in NAPI doesn't
benefit performance in anyway. Just get rid of it.
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be.h | 2 -
drivers/net/ethernet/emulex/benet/be_main.c | 30 +++++++++-----------------
2 files changed, 11 insertions(+), 21 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 5ccfce7..d952ca2 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -114,7 +114,6 @@ static inline char *nic_name(struct pci_dev *pdev)
#define MAX_ROCE_EQS 5
#define MAX_MSIX_VECTORS 32
#define MIN_MSIX_VECTORS 1
-#define BE_TX_BUDGET 256
#define BE_NAPI_WEIGHT 64
#define MAX_RX_POST BE_NAPI_WEIGHT /* Frags posted at a time */
#define RX_FRAGS_REFILL_WM (RX_Q_LEN - MAX_RX_POST)
@@ -200,7 +199,6 @@ struct be_eq_obj {
u8 idx; /* array index */
u8 msix_idx;
- u16 tx_budget;
u16 spurious_intr;
struct napi_struct napi;
struct be_adapter *adapter;
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 04a8092..53727ad 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -2143,7 +2143,6 @@ static int be_evt_queues_create(struct be_adapter *adapter)
napi_hash_add(&eqo->napi);
aic = &adapter->aic_obj[i];
eqo->adapter = adapter;
- eqo->tx_budget = BE_TX_BUDGET;
eqo->idx = i;
aic->max_eqd = BE_MAX_EQD;
aic->enable = true;
@@ -2459,20 +2458,19 @@ static inline void lancer_update_tx_err(struct be_tx_obj *txo, u32 status)
}
}
-static bool be_process_tx(struct be_adapter *adapter, struct be_tx_obj *txo,
- int budget, int idx)
+static void be_process_tx(struct be_adapter *adapter, struct be_tx_obj *txo,
+ int idx)
{
struct be_eth_tx_compl *txcp;
- int num_wrbs = 0, work_done;
+ int num_wrbs = 0, work_done = 0;
u32 compl_status;
+ u16 last_idx;
+
+ while ((txcp = be_tx_compl_get(&txo->cq))) {
+ last_idx = GET_TX_COMPL_BITS(wrb_index, txcp);
+ num_wrbs += be_tx_compl_process(adapter, txo, last_idx);
+ work_done++;
- for (work_done = 0; work_done < budget; work_done++) {
- txcp = be_tx_compl_get(&txo->cq);
- if (!txcp)
- break;
- num_wrbs += be_tx_compl_process(adapter, txo,
- GET_TX_COMPL_BITS(wrb_index,
- txcp));
compl_status = GET_TX_COMPL_BITS(status, txcp);
if (compl_status) {
if (lancer_chip(adapter))
@@ -2497,7 +2495,6 @@ static bool be_process_tx(struct be_adapter *adapter, struct be_tx_obj *txo,
tx_stats(txo)->tx_compl += work_done;
u64_stats_update_end(&tx_stats(txo)->sync_compl);
}
- return (work_done < budget); /* Done */
}
int be_poll(struct napi_struct *napi, int budget)
@@ -2506,17 +2503,12 @@ int be_poll(struct napi_struct *napi, int budget)
struct be_adapter *adapter = eqo->adapter;
int max_work = 0, work, i, num_evts;
struct be_rx_obj *rxo;
- bool tx_done;
num_evts = events_get(eqo);
/* Process all TXQs serviced by this EQ */
- for (i = eqo->idx; i < adapter->num_tx_qs; i += adapter->num_evt_qs) {
- tx_done = be_process_tx(adapter, &adapter->tx_obj[i],
- eqo->tx_budget, i);
- if (!tx_done)
- max_work = budget;
- }
+ for (i = eqo->idx; i < adapter->num_tx_qs; i += adapter->num_evt_qs)
+ be_process_tx(adapter, &adapter->tx_obj[i], i);
if (be_lock_napi(eqo)) {
/* This loop will iterate twice for EQ0 in which
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 07/12] be2net: remove unncessary gotos
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Kalesh AP <kalesh.purayil@emulex.com>
In cases where there is no extra code to handle an error, this patch replaces
gotos with a direct return statement.
Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be_main.c | 16 ++++++----------
1 files changed, 6 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 7e89e5b..cdad8bc 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -4090,18 +4090,15 @@ static int lancer_fw_download(struct be_adapter *adapter,
if (!IS_ALIGNED(fw->size, sizeof(u32))) {
dev_err(dev, "FW image size should be multiple of 4\n");
- status = -EINVAL;
- goto lancer_fw_exit;
+ return -EINVAL;
}
flash_cmd.size = sizeof(struct lancer_cmd_req_write_object)
+ LANCER_FW_DOWNLOAD_CHUNK;
flash_cmd.va = dma_alloc_coherent(dev, flash_cmd.size,
&flash_cmd.dma, GFP_KERNEL);
- if (!flash_cmd.va) {
- status = -ENOMEM;
- goto lancer_fw_exit;
- }
+ if (!flash_cmd.va)
+ return -ENOMEM;
dest_image_ptr = flash_cmd.va +
sizeof(struct lancer_cmd_req_write_object);
@@ -4139,7 +4136,7 @@ static int lancer_fw_download(struct be_adapter *adapter,
dma_free_coherent(dev, flash_cmd.size, flash_cmd.va, flash_cmd.dma);
if (status) {
dev_err(dev, "Firmware load error\n");
- goto lancer_fw_exit;
+ return be_cmd_status(status);
}
dev_info(dev, "Firmware flashed successfully\n");
@@ -4151,13 +4148,12 @@ static int lancer_fw_download(struct be_adapter *adapter,
if (status) {
dev_err(dev, "Adapter busy, could not reset FW\n");
dev_err(dev, "Reboot server to activate new FW\n");
- goto lancer_fw_exit;
}
} else if (change_status != LANCER_NO_RESET_NEEDED) {
dev_info(dev, "Reboot server to activate new FW\n");
}
-lancer_fw_exit:
- return status;
+
+ return 0;
}
#define UFI_TYPE2 2
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 05/12] be2net: Add a dma_mapping_error counter in ethtool
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
From: Vasundhara Volam <vasundhara.volam@emulex.com>
Add a dma_mapping_error counter to count the number of packets dropped
due to DMA mapping errors.
Signed-off-by: Vasundhara Volam <vasundhara.volam@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be.h | 1 +
drivers/net/ethernet/emulex/benet/be_ethtool.c | 2 ++
drivers/net/ethernet/emulex/benet/be_main.c | 3 ++-
3 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index d491ac6..069bdda 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -323,6 +323,7 @@ struct be_rx_obj {
struct be_drv_stats {
u32 be_on_die_temperature;
u32 eth_red_drops;
+ u32 dma_map_errors;
u32 rx_drops_no_pbuf;
u32 rx_drops_no_txpb;
u32 rx_drops_no_erx_descr;
diff --git a/drivers/net/ethernet/emulex/benet/be_ethtool.c b/drivers/net/ethernet/emulex/benet/be_ethtool.c
index 3a7ade4..2fd3826 100644
--- a/drivers/net/ethernet/emulex/benet/be_ethtool.c
+++ b/drivers/net/ethernet/emulex/benet/be_ethtool.c
@@ -119,6 +119,8 @@ static const struct be_ethtool_stat et_stats[] = {
* is more than 9018 bytes
*/
{DRVSTAT_INFO(rx_drops_mtu)},
+ /* Number of dma mapping errors */
+ {DRVSTAT_INFO(dma_map_errors)},
/* Number of packets dropped due to random early drop function */
{DRVSTAT_INFO(eth_red_drops)},
{DRVSTAT_INFO(be_on_die_temperature)},
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index e9658f7..c95998f 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -849,6 +849,7 @@ dma_err:
unmap_tx_frag(dev, wrb, map_single);
map_single = false;
copied -= wrb->frag_len;
+ adapter->drv_stats.dma_map_errors++;
queue_head_inc(txq);
}
return 0;
@@ -1877,7 +1878,7 @@ static void be_post_rx_frags(struct be_rx_obj *rxo, gfp_t gfp)
if (dma_mapping_error(dev, page_dmaaddr)) {
put_page(pagep);
pagep = NULL;
- rx_stats(rxo)->rx_post_fail++;
+ adapter->drv_stats.dma_map_errors++;
break;
}
page_offset = 0;
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 02/12] be2net: shorten AMAP_GET/SET_BITS() macro calls
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
The AMAP_GET/SET_BITS() macro calls take structure name as a parameter
and hence are long and span more than one line. Replace these calls
with a wrapper macros for RX/Tx compls and TX wrb. This results in fewer
lines and more readable code in be_main.c
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be.h | 12 +++
drivers/net/ethernet/emulex/benet/be_main.c | 115 +++++++++++----------------
2 files changed, 58 insertions(+), 69 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 43e08d0..d9cd885 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -661,6 +661,18 @@ static inline u32 amap_get(void *ptr, u32 dw_offset, u32 mask, u32 offset)
amap_mask(sizeof(((_struct *)0)->field)), \
AMAP_BIT_OFFSET(_struct, field))
+#define GET_RX_COMPL_V0_BITS(field, ptr) \
+ AMAP_GET_BITS(struct amap_eth_rx_compl_v0, field, ptr)
+
+#define GET_RX_COMPL_V1_BITS(field, ptr) \
+ AMAP_GET_BITS(struct amap_eth_rx_compl_v1, field, ptr)
+
+#define GET_TX_COMPL_BITS(field, ptr) \
+ AMAP_GET_BITS(struct amap_eth_tx_compl, field, ptr)
+
+#define SET_TX_WRB_HDR_BITS(field, ptr, val) \
+ AMAP_SET_BITS(struct amap_eth_hdr_wrb, field, ptr, val)
+
#define be_dws_cpu_to_le(wrb, len) swap_dws(wrb, len)
#define be_dws_le_to_cpu(wrb, len) swap_dws(wrb, len)
static inline void swap_dws(void *wrb, int len)
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 3d33049..810ce5c 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -738,38 +738,37 @@ static void wrb_fill_hdr(struct be_adapter *adapter, struct be_eth_hdr_wrb *hdr,
memset(hdr, 0, sizeof(*hdr));
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, crc, hdr, 1);
+ SET_TX_WRB_HDR_BITS(crc, hdr, 1);
if (skb_is_gso(skb)) {
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, lso, hdr, 1);
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, lso_mss,
- hdr, skb_shinfo(skb)->gso_size);
+ SET_TX_WRB_HDR_BITS(lso, hdr, 1);
+ SET_TX_WRB_HDR_BITS(lso_mss, hdr, skb_shinfo(skb)->gso_size);
if (skb_is_gso_v6(skb) && !lancer_chip(adapter))
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, lso6, hdr, 1);
+ SET_TX_WRB_HDR_BITS(lso6, hdr, 1);
} else if (skb->ip_summed == CHECKSUM_PARTIAL) {
if (skb->encapsulation) {
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, ipcs, hdr, 1);
+ SET_TX_WRB_HDR_BITS(ipcs, hdr, 1);
proto = skb_inner_ip_proto(skb);
} else {
proto = skb_ip_proto(skb);
}
if (proto == IPPROTO_TCP)
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, tcpcs, hdr, 1);
+ SET_TX_WRB_HDR_BITS(tcpcs, hdr, 1);
else if (proto == IPPROTO_UDP)
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, udpcs, hdr, 1);
+ SET_TX_WRB_HDR_BITS(udpcs, hdr, 1);
}
if (vlan_tx_tag_present(skb)) {
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, vlan, hdr, 1);
+ SET_TX_WRB_HDR_BITS(vlan, hdr, 1);
vlan_tag = be_get_tx_vlan_tag(adapter, skb);
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, vlan_tag, hdr, vlan_tag);
+ SET_TX_WRB_HDR_BITS(vlan_tag, hdr, vlan_tag);
}
/* To skip HW VLAN tagging: evt = 1, compl = 0 */
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, complete, hdr, !skip_hw_vlan);
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, event, hdr, 1);
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, num_wrb, hdr, wrb_cnt);
- AMAP_SET_BITS(struct amap_eth_hdr_wrb, len, hdr, len);
+ SET_TX_WRB_HDR_BITS(complete, hdr, !skip_hw_vlan);
+ SET_TX_WRB_HDR_BITS(event, hdr, 1);
+ SET_TX_WRB_HDR_BITS(num_wrb, hdr, wrb_cnt);
+ SET_TX_WRB_HDR_BITS(len, hdr, len);
}
static void unmap_tx_frag(struct device *dev, struct be_eth_wrb *wrb,
@@ -1753,65 +1752,46 @@ static void be_rx_compl_process_gro(struct be_rx_obj *rxo,
static void be_parse_rx_compl_v1(struct be_eth_rx_compl *compl,
struct be_rx_compl_info *rxcp)
{
- rxcp->pkt_size =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, pktsize, compl);
- rxcp->vlanf = AMAP_GET_BITS(struct amap_eth_rx_compl_v1, vtp, compl);
- rxcp->err = AMAP_GET_BITS(struct amap_eth_rx_compl_v1, err, compl);
- rxcp->tcpf = AMAP_GET_BITS(struct amap_eth_rx_compl_v1, tcpf, compl);
- rxcp->udpf = AMAP_GET_BITS(struct amap_eth_rx_compl_v1, udpf, compl);
- rxcp->ip_csum =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, ipcksm, compl);
- rxcp->l4_csum =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, l4_cksm, compl);
- rxcp->ipv6 =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, ip_version, compl);
- rxcp->num_rcvd =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, numfrags, compl);
- rxcp->pkt_type =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, cast_enc, compl);
- rxcp->rss_hash =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, rsshash, compl);
+ rxcp->pkt_size = GET_RX_COMPL_V1_BITS(pktsize, compl);
+ rxcp->vlanf = GET_RX_COMPL_V1_BITS(vtp, compl);
+ rxcp->err = GET_RX_COMPL_V1_BITS(err, compl);
+ rxcp->tcpf = GET_RX_COMPL_V1_BITS(tcpf, compl);
+ rxcp->udpf = GET_RX_COMPL_V1_BITS(udpf, compl);
+ rxcp->ip_csum = GET_RX_COMPL_V1_BITS(ipcksm, compl);
+ rxcp->l4_csum = GET_RX_COMPL_V1_BITS(l4_cksm, compl);
+ rxcp->ipv6 = GET_RX_COMPL_V1_BITS(ip_version, compl);
+ rxcp->num_rcvd = GET_RX_COMPL_V1_BITS(numfrags, compl);
+ rxcp->pkt_type = GET_RX_COMPL_V1_BITS(cast_enc, compl);
+ rxcp->rss_hash = GET_RX_COMPL_V1_BITS(rsshash, compl);
if (rxcp->vlanf) {
- rxcp->qnq = AMAP_GET_BITS(struct amap_eth_rx_compl_v1, qnq,
- compl);
- rxcp->vlan_tag = AMAP_GET_BITS(struct amap_eth_rx_compl_v1,
- vlan_tag, compl);
+ rxcp->qnq = GET_RX_COMPL_V1_BITS(qnq, compl);
+ rxcp->vlan_tag = GET_RX_COMPL_V1_BITS(vlan_tag, compl);
}
- rxcp->port = AMAP_GET_BITS(struct amap_eth_rx_compl_v1, port, compl);
+ rxcp->port = GET_RX_COMPL_V1_BITS(port, compl);
rxcp->tunneled =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v1, tunneled, compl);
+ GET_RX_COMPL_V1_BITS(tunneled, compl);
}
static void be_parse_rx_compl_v0(struct be_eth_rx_compl *compl,
struct be_rx_compl_info *rxcp)
{
- rxcp->pkt_size =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, pktsize, compl);
- rxcp->vlanf = AMAP_GET_BITS(struct amap_eth_rx_compl_v0, vtp, compl);
- rxcp->err = AMAP_GET_BITS(struct amap_eth_rx_compl_v0, err, compl);
- rxcp->tcpf = AMAP_GET_BITS(struct amap_eth_rx_compl_v0, tcpf, compl);
- rxcp->udpf = AMAP_GET_BITS(struct amap_eth_rx_compl_v0, udpf, compl);
- rxcp->ip_csum =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, ipcksm, compl);
- rxcp->l4_csum =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, l4_cksm, compl);
- rxcp->ipv6 =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, ip_version, compl);
- rxcp->num_rcvd =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, numfrags, compl);
- rxcp->pkt_type =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, cast_enc, compl);
- rxcp->rss_hash =
- AMAP_GET_BITS(struct amap_eth_rx_compl_v0, rsshash, compl);
+ rxcp->pkt_size = GET_RX_COMPL_V0_BITS(pktsize, compl);
+ rxcp->vlanf = GET_RX_COMPL_V0_BITS(vtp, compl);
+ rxcp->err = GET_RX_COMPL_V0_BITS(err, compl);
+ rxcp->tcpf = GET_RX_COMPL_V0_BITS(tcpf, compl);
+ rxcp->udpf = GET_RX_COMPL_V0_BITS(udpf, compl);
+ rxcp->ip_csum = GET_RX_COMPL_V0_BITS(ipcksm, compl);
+ rxcp->l4_csum = GET_RX_COMPL_V0_BITS(l4_cksm, compl);
+ rxcp->ipv6 = GET_RX_COMPL_V0_BITS(ip_version, compl);
+ rxcp->num_rcvd = GET_RX_COMPL_V0_BITS(numfrags, compl);
+ rxcp->pkt_type = GET_RX_COMPL_V0_BITS(cast_enc, compl);
+ rxcp->rss_hash = GET_RX_COMPL_V0_BITS(rsshash, compl);
if (rxcp->vlanf) {
- rxcp->qnq = AMAP_GET_BITS(struct amap_eth_rx_compl_v0, qnq,
- compl);
- rxcp->vlan_tag = AMAP_GET_BITS(struct amap_eth_rx_compl_v0,
- vlan_tag, compl);
+ rxcp->qnq = GET_RX_COMPL_V0_BITS(qnq, compl);
+ rxcp->vlan_tag = GET_RX_COMPL_V0_BITS(vlan_tag, compl);
}
- rxcp->port = AMAP_GET_BITS(struct amap_eth_rx_compl_v0, port, compl);
- rxcp->ip_frag = AMAP_GET_BITS(struct amap_eth_rx_compl_v0,
- ip_frag, compl);
+ rxcp->port = GET_RX_COMPL_V0_BITS(port, compl);
+ rxcp->ip_frag = GET_RX_COMPL_V0_BITS(ip_frag, compl);
}
static struct be_rx_compl_info *be_rx_compl_get(struct be_rx_obj *rxo)
@@ -2091,9 +2071,7 @@ static void be_tx_compl_clean(struct be_adapter *adapter)
num_wrbs = 0;
txq = &txo->q;
while ((txcp = be_tx_compl_get(&txo->cq))) {
- end_idx =
- AMAP_GET_BITS(struct amap_eth_tx_compl,
- wrb_index, txcp);
+ end_idx = GET_TX_COMPL_BITS(wrb_index, txcp);
num_wrbs += be_tx_compl_process(adapter, txo,
end_idx);
cmpl++;
@@ -2454,9 +2432,8 @@ static bool be_process_tx(struct be_adapter *adapter, struct be_tx_obj *txo,
if (!txcp)
break;
num_wrbs += be_tx_compl_process(adapter, txo,
- AMAP_GET_BITS(struct
- amap_eth_tx_compl,
- wrb_index, txcp));
+ GET_TX_COMPL_BITS(wrb_index,
+ txcp));
}
if (work_done) {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 03/12] be2net: add a description for counter rx_input_fifo_overflow_drop
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
In-Reply-To: <1409552664-15990-1-git-send-email-sathya.perla@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
drivers/net/ethernet/emulex/benet/be_ethtool.c | 5 +++++
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_ethtool.c b/drivers/net/ethernet/emulex/benet/be_ethtool.c
index 0cd3311..d8d7a4a 100644
--- a/drivers/net/ethernet/emulex/benet/be_ethtool.c
+++ b/drivers/net/ethernet/emulex/benet/be_ethtool.c
@@ -78,6 +78,11 @@ static const struct be_ethtool_stat et_stats[] = {
* fifo must never overflow.
*/
{DRVSTAT_INFO(rxpp_fifo_overflow_drop)},
+ /* Received packets dropped when the RX block runs out of space in
+ * one of its input FIFOs. This could happen due a long burst of
+ * minimum-sized (64b) frames in the receive path.
+ * This counter may also be erroneously incremented rarely.
+ */
{DRVSTAT_INFO(rx_input_fifo_overflow_drop)},
{DRVSTAT_INFO(rx_ip_checksum_errs)},
{DRVSTAT_INFO(rx_tcp_checksum_errs)},
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 00/12] be2net: patch set
From: Sathya Perla @ 2014-09-01 6:24 UTC (permalink / raw)
To: netdev
Patch 1 adds a few new log messages to help debugging in failure cases.
Patch 2 uses new macros for parsing RX/TX completions and TX wrbs to
help shorten the lines.
Patch 3 adds a description for the RX counter rx_input_fifo_overflow_drop.
Patch 4 adds TX completion error statistics reporting via ethtool.
Patch 5 adds a dma_mapping_error counter and its reporting via ethtool.
Patch 6 fixes up log messages in the Lancer FW download path.
Patch 7 replaces gotos with direct return statements.
Patch 8 cleans up be_change_mtu() code by using a new macro BE_MAX_MTU
Patch 9 makes be_cmd_get_regs() routine to return an integer status
similar to other FW cmd routines in be_cmds.c
Patch 10 gets rid of TX budget as enforcing a budget on TX completion
processing in NAPI is neither suggested nor it provides a performance benefit.
Patch 11 defines and uses a new macro for_all_tx_queues_on_eq() similar
to the RX processing code.
Patch 12 queries max_tx_qs from the FW for BE3 super-nic profiles.
For those profiles, the driver cannot assume a constant BE3_MAX_TX_QS value,
as the value may change for each function.
Please consider applying this patch set to the net-next tree. Thanks!
Kalesh AP (4):
be2net: Add TX completion error statistics in ethtool
be2net: fix log messages in lancer FW download path
be2net: remove unncessary gotos
be2net: define BE_MAX_MTU
Sathya Perla (5):
be2net: add a few log messages
be2net: shorten AMAP_GET/SET_BITS() macro calls
be2net: add a description for counter rx_input_fifo_overflow_drop
be2net: get rid of TX budget
be2net: define macro for_all_tx_queues_on_eq()
Suresh Reddy (1):
be2net: query max_tx_qs for BE3 super-nic profile from FW
Vasundhara Volam (2):
be2net: Add a dma_mapping_error counter in ethtool
be2net: make be_cmd_get_regs() return a status
drivers/net/ethernet/emulex/benet/be.h | 28 +++-
drivers/net/ethernet/emulex/benet/be_cmds.c | 14 +-
drivers/net/ethernet/emulex/benet/be_cmds.h | 2 +-
drivers/net/ethernet/emulex/benet/be_ethtool.c | 35 +++
drivers/net/ethernet/emulex/benet/be_hw.h | 12 +
drivers/net/ethernet/emulex/benet/be_main.c | 282 +++++++++++++-----------
6 files changed, 232 insertions(+), 141 deletions(-)
^ permalink raw reply
* Re: [PATCH (net.git) 0/4 (v2)] stmmac EEE fixes
From: David Miller @ 2014-09-01 5:01 UTC (permalink / raw)
To: peppe.cavallaro; +Cc: netdev
In-Reply-To: <1409206304-23239-1-git-send-email-peppe.cavallaro@st.com>
From: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Date: Thu, 28 Aug 2014 08:11:40 +0200
> This is a subset of patches to provide some fixes for the EEE support inside the
> driver.
> Patches have been tested on boards EEE capable plugged on switch w/ w/o EEE
> support.
Series applied, thanks.
^ permalink raw reply
* Re: [PATCH net-next v2] r8152: reduce the number of Tx
From: David Miller @ 2014-09-01 4:38 UTC (permalink / raw)
To: hayeswang; +Cc: netdev, nic_swsd, linux-kernel, linux-usb
In-Reply-To: <0835B3720019904CB8F7AA43166CEEB2516A75@RTITMBSV03.realtek.com.tw>
From: Hayes Wang <hayeswang@realtek.com>
Date: Mon, 1 Sep 2014 02:46:38 +0000
> From: David Miller [mailto:davem@davemloft.net]
> [...]
>> This driver has a loop that iterates MAX_TX times to initialize both
>> the RX and TX buffers.
>>
>> So if they are not equal, it can't possibly work.
>
> Excuse me. I don't find that the "RTL8152_MAX_TX" is used by
> rx buffer. The rx buffer only use "RTL8152_MAX_RX". Although
> the tx buffer size is determined by "rx_buf_sz", the tx buffers
> are independent of rx buffers.
Ok, but that is really confusing code.
^ permalink raw reply
* RE: [PATCH net-next v2] r8152: reduce the number of Tx
From: Hayes Wang @ 2014-09-01 2:46 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, nic_swsd, linux-kernel@vger.kernel.org,
linux-usb@vger.kernel.org
In-Reply-To: <20140829.202504.655684636693672249.davem@davemloft.net>
From: David Miller [mailto:davem@davemloft.net]
[...]
> This driver has a loop that iterates MAX_TX times to initialize both
> the RX and TX buffers.
>
> So if they are not equal, it can't possibly work.
Excuse me. I don't find that the "RTL8152_MAX_TX" is used by
rx buffer. The rx buffer only use "RTL8152_MAX_RX". Although
the tx buffer size is determined by "rx_buf_sz", the tx buffers
are independent of rx buffers.
Best Regards,
Hayes
^ permalink raw reply
* Re: [PATCH 0/2] Get rid of ndo_xmit_flush
From: Rusty Russell @ 2014-09-01 0:37 UTC (permalink / raw)
To: Jesper Dangaard Brouer
Cc: brouer, David Miller, netdev, therbert, jhs, hannes, edumazet,
jeffrey.t.kirsher, dborkman
In-Reply-To: <20140826121347.0ec7f2ac@redhat.com>
Jesper Dangaard Brouer <brouer@redhat.com> writes:
> On Tue, 26 Aug 2014 08:28:15 +0200 Jesper Dangaard Brouer <brouer@redhat.com> wrote:
>> On Mon, 25 Aug 2014 16:34:58 -0700 (PDT) David Miller <davem@davemloft.net> wrote:
>>
>> > Given Jesper's performance numbers, it's not the way to go.
>> >
>> > Instead, go with a signalling scheme via new boolean skb->xmit_more.
>>
>> I'll do benchmarking based on this new API proposal today.
>
> While establish an accurate baseline for my measurements. I'm
> starting to see too much variation in my trafgen measurements.
> Meaning that we unfortunately cannot use it to measure variations on
> the nanosec scale.
>
> I'm measuring the packets per sec via "ifpps", and calculating an
> average over the measurements, via the following oneliner:
>
> $ ifpps -clod eth5 -t 1000 | awk 'BEGIN{txsum=0; rxsum=0; n=0} /[[:digit:]]/ {txsum+=$11;rxsum+=$3;n++; printf "instant rx:%u tx:%u pps n:%u average: rx:%d tx:%d pps\n", $3, $11, n, rxsum/n, txsum/n }'
FYI, this is what I use for this kind of thing:
https://github.com/rustyrussell/stats
>From the README:
This filter finds identical lines and collapses all the numbers into a
range, average and standard deviation; it can also print out all the
numbers in CSV form for import into spreadsheets, etc.
Cheers,
Rusty.
^ permalink raw reply
* [PATCH net-next] sock: deduplicate errqueue dequeue
From: Willem de Bruijn @ 2014-09-01 1:30 UTC (permalink / raw)
To: netdev; +Cc: davem, Willem de Bruijn
sk->sk_error_queue is dequeued in four locations. All share the
exact same logic. Deduplicate.
Also collapse the two critical sections for dequeue (at the top of
the recv handler) and signal (at the bottom).
This moves signal generation for the next packet forward, which should
be harmless.
It also changes the behavior if the recv handler exits early with an
error. Previously, a signal for follow-up packets on the errqueue
would then not be scheduled. The new behavior, to always signal, is
arguably a bug fix.
For rxrpc, the change causes the same function to be called repeatedly
for each queued packet (because the recv handler == sk_error_report).
It is likely that all packets will fail for the same reason (e.g.,
memory exhaustion).
This code runs without sk_lock held, so it is not safe to trust that
sk->sk_err is immutable inbetween releasing q->lock and the subsequent
test. Introduce int err just to avoid this potential race.
Signed-off-by: Willem de Bruijn <willemb@google.com>
---
include/net/sock.h | 1 +
net/core/skbuff.c | 20 ++++++++++++++++++++
net/core/sock.c | 14 ++------------
net/ipv4/ip_sockglue.c | 15 ++-------------
net/ipv6/datagram.c | 15 ++-------------
net/rxrpc/ar-error.c | 14 +-------------
6 files changed, 28 insertions(+), 51 deletions(-)
diff --git a/include/net/sock.h b/include/net/sock.h
index 7f2ab72..3fde613 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -2041,6 +2041,7 @@ void sk_stop_timer(struct sock *sk, struct timer_list *timer);
int sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb);
+struct sk_buff *sock_dequeue_err_skb(struct sock *sk);
/*
* Recover an error report and clear atomically
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 163b673..53ce536 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -3491,6 +3491,26 @@ int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb)
}
EXPORT_SYMBOL(sock_queue_err_skb);
+struct sk_buff *sock_dequeue_err_skb(struct sock *sk)
+{
+ struct sk_buff_head *q = &sk->sk_error_queue;
+ struct sk_buff *skb, *skb_next;
+ int err = 0;
+
+ spin_lock_bh(&q->lock);
+ skb = __skb_dequeue(q);
+ if (skb && (skb_next = skb_peek(q)))
+ err = SKB_EXT_ERR(skb_next)->ee.ee_errno;
+ spin_unlock_bh(&q->lock);
+
+ sk->sk_err = err;
+ if (err)
+ sk->sk_error_report(sk);
+
+ return skb;
+}
+EXPORT_SYMBOL(sock_dequeue_err_skb);
+
void __skb_tstamp_tx(struct sk_buff *orig_skb,
struct skb_shared_hwtstamps *hwtstamps,
struct sock *sk, int tstype)
diff --git a/net/core/sock.c b/net/core/sock.c
index f7f2352..f1a638e 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -2488,11 +2488,11 @@ int sock_recv_errqueue(struct sock *sk, struct msghdr *msg, int len,
int level, int type)
{
struct sock_exterr_skb *serr;
- struct sk_buff *skb, *skb2;
+ struct sk_buff *skb;
int copied, err;
err = -EAGAIN;
- skb = skb_dequeue(&sk->sk_error_queue);
+ skb = sock_dequeue_err_skb(sk);
if (skb == NULL)
goto out;
@@ -2513,16 +2513,6 @@ int sock_recv_errqueue(struct sock *sk, struct msghdr *msg, int len,
msg->msg_flags |= MSG_ERRQUEUE;
err = copied;
- /* Reset and regenerate socket error */
- spin_lock_bh(&sk->sk_error_queue.lock);
- sk->sk_err = 0;
- if ((skb2 = skb_peek(&sk->sk_error_queue)) != NULL) {
- sk->sk_err = SKB_EXT_ERR(skb2)->ee.ee_errno;
- spin_unlock_bh(&sk->sk_error_queue.lock);
- sk->sk_error_report(sk);
- } else
- spin_unlock_bh(&sk->sk_error_queue.lock);
-
out_free_skb:
kfree_skb(skb);
out:
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index 5cb830c..455e75b 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -405,7 +405,7 @@ void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 port, u32 inf
int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
{
struct sock_exterr_skb *serr;
- struct sk_buff *skb, *skb2;
+ struct sk_buff *skb;
DECLARE_SOCKADDR(struct sockaddr_in *, sin, msg->msg_name);
struct {
struct sock_extended_err ee;
@@ -415,7 +415,7 @@ int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
int copied;
err = -EAGAIN;
- skb = skb_dequeue(&sk->sk_error_queue);
+ skb = sock_dequeue_err_skb(sk);
if (skb == NULL)
goto out;
@@ -462,17 +462,6 @@ int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
msg->msg_flags |= MSG_ERRQUEUE;
err = copied;
- /* Reset and regenerate socket error */
- spin_lock_bh(&sk->sk_error_queue.lock);
- sk->sk_err = 0;
- skb2 = skb_peek(&sk->sk_error_queue);
- if (skb2 != NULL) {
- sk->sk_err = SKB_EXT_ERR(skb2)->ee.ee_errno;
- spin_unlock_bh(&sk->sk_error_queue.lock);
- sk->sk_error_report(sk);
- } else
- spin_unlock_bh(&sk->sk_error_queue.lock);
-
out_free_skb:
kfree_skb(skb);
out:
diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c
index 1844e87..2cdc383 100644
--- a/net/ipv6/datagram.c
+++ b/net/ipv6/datagram.c
@@ -332,7 +332,7 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
{
struct ipv6_pinfo *np = inet6_sk(sk);
struct sock_exterr_skb *serr;
- struct sk_buff *skb, *skb2;
+ struct sk_buff *skb;
DECLARE_SOCKADDR(struct sockaddr_in6 *, sin, msg->msg_name);
struct {
struct sock_extended_err ee;
@@ -342,7 +342,7 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
int copied;
err = -EAGAIN;
- skb = skb_dequeue(&sk->sk_error_queue);
+ skb = sock_dequeue_err_skb(sk);
if (skb == NULL)
goto out;
@@ -415,17 +415,6 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
msg->msg_flags |= MSG_ERRQUEUE;
err = copied;
- /* Reset and regenerate socket error */
- spin_lock_bh(&sk->sk_error_queue.lock);
- sk->sk_err = 0;
- if ((skb2 = skb_peek(&sk->sk_error_queue)) != NULL) {
- sk->sk_err = SKB_EXT_ERR(skb2)->ee.ee_errno;
- spin_unlock_bh(&sk->sk_error_queue.lock);
- sk->sk_error_report(sk);
- } else {
- spin_unlock_bh(&sk->sk_error_queue.lock);
- }
-
out_free_skb:
kfree_skb(skb);
out:
diff --git a/net/rxrpc/ar-error.c b/net/rxrpc/ar-error.c
index db57458..74c0fcd 100644
--- a/net/rxrpc/ar-error.c
+++ b/net/rxrpc/ar-error.c
@@ -37,7 +37,7 @@ void rxrpc_UDP_error_report(struct sock *sk)
_enter("%p{%d}", sk, local->debug_id);
- skb = skb_dequeue(&sk->sk_error_queue);
+ skb = sock_dequeue_err_skb(sk);
if (!skb) {
_leave("UDP socket errqueue empty");
return;
@@ -111,18 +111,6 @@ void rxrpc_UDP_error_report(struct sock *sk)
skb_queue_tail(&trans->error_queue, skb);
rxrpc_queue_work(&trans->error_handler);
- /* reset and regenerate socket error */
- spin_lock_bh(&sk->sk_error_queue.lock);
- sk->sk_err = 0;
- skb = skb_peek(&sk->sk_error_queue);
- if (skb) {
- sk->sk_err = SKB_EXT_ERR(skb)->ee.ee_errno;
- spin_unlock_bh(&sk->sk_error_queue.lock);
- sk->sk_error_report(sk);
- } else {
- spin_unlock_bh(&sk->sk_error_queue.lock);
- }
-
_leave("");
}
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next] sock: consistent errqueue errors and signals
From: Willem de Bruijn @ 2014-09-01 1:28 UTC (permalink / raw)
To: netdev; +Cc: davem, Willem de Bruijn
When a socket error is pending, send()/recv() must abort their normal
operation and return the error. An error means having non-zero
sk->sk_err or having non-empty sk->sk_error_queue.
Currently, the behavior for the second is inconsistent depending on
whether an error has previously been dequeued. In all cases,
recv()/send() test sk->sk_err. This is not modified on enqueue onto
the error queue, so may be 0. It is modified on dequeue, however, to
match the queued skb's errno. I observed the following when two errors
were queued:
ret = poll(pollfd, 1, -1);
assert(ret == 1);
assert(pollfd.revents == POLLERR);
ret = recv(fd, buf, size, MSG_NONBLOCK);
assert(ret == -1 && errno == EAGAIN); /* <-- A */
ret = recv(fd, buf, size, MSG_ERRQUEUE);
assert(ret > 0);
ret = recv(fd, buf, size, MSG_NONBLOCK);
assert(ret == -1 && errno == ENOMSG); /* <-- B */
ret = recv(fd, buf, size, MSG_ERRQUEUE);
assert(ret > 0);
The recv call in B returns the error code embedded in
SKB_EXT_ERR(skb), in this case ENOMSG, because I am working with
timestamps. The recv call in A should have returned the
same.
Implement this behavior. This may surprise existing applications.
Also make the wake-up signal when data is ready on the error queue
consistent between enqueue and dequeue: use sk_error_report in both
cases.
Signed-off-by: Willem de Bruijn <willemb@google.com>
---
This approach leaves one issue:
The states of sk->sk_err and sk->sk_error_queue are related, but only
loosely. Error queue enqueue, dequeue and other code may overwrite
sk->sk_err unconditionally. For one, sock_error will reset
sk->sk_err to 0 even if sk->sk_error_queue is not empty. If socket
calls should abort on all errors, then should be change to test
sk_error_queue.qlen. But, doing so requires taking a lock in a busy
data path.
---
net/core/skbuff.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 163b673..f7a280b 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -3485,8 +3485,11 @@ int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb)
skb_dst_force(skb);
skb_queue_tail(&sk->sk_error_queue, skb);
+ sk->sk_err = SKB_EXT_ERR(skb)->ee.ee_errno;
+
if (!sock_flag(sk, SOCK_DEAD))
- sk->sk_data_ready(sk);
+ sk->sk_error_report(sk);
+
return 0;
}
EXPORT_SYMBOL(sock_queue_err_skb);
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next] net-timestamp: expand documentation
From: Willem de Bruijn @ 2014-09-01 1:27 UTC (permalink / raw)
To: netdev; +Cc: davem, Willem de Bruijn
Expand Documentation/networking/timestamping.txt with new
interfaces and bytestream timestamping. Also minor
cleanup of the other text.
Import txtimestamp.c test of the new features.
Signed-off-by: Willem de Bruijn <willemb@google.com>
---
Documentation/networking/timestamping.txt | 368 ++++++++++++----
Documentation/networking/timestamping/Makefile | 10 +-
.../networking/timestamping/txtimestamp.c | 470 +++++++++++++++++++++
3 files changed, 764 insertions(+), 84 deletions(-)
create mode 100644 Documentation/networking/timestamping/txtimestamp.c
diff --git a/Documentation/networking/timestamping.txt b/Documentation/networking/timestamping.txt
index 897f942..412f45c 100644
--- a/Documentation/networking/timestamping.txt
+++ b/Documentation/networking/timestamping.txt
@@ -1,102 +1,307 @@
-The existing interfaces for getting network packages time stamped are:
+
+1. Control Interfaces
+
+The interfaces for receiving network packages timestamps are:
* SO_TIMESTAMP
- Generate time stamp for each incoming packet using the (not necessarily
- monotonous!) system time. Result is returned via recv_msg() in a
- control message as timeval (usec resolution).
+ Generates a timestamp for each incoming packet in (not necessarily
+ monotonic) system time. Reports the timestamp via recvmsg() in a
+ control message as struct timeval (usec resolution).
* SO_TIMESTAMPNS
- Same time stamping mechanism as SO_TIMESTAMP, but returns result as
- timespec (nsec resolution).
+ Same timestamping mechanism as SO_TIMESTAMP, but reports the
+ timestamp as struct timespec (nsec resolution).
* IP_MULTICAST_LOOP + SO_TIMESTAMP[NS]
- Only for multicasts: approximate send time stamp by receiving the looped
- packet and using its receive time stamp.
+ Only for multicast:approximate transmit timestamp obtained by
+ reading the looped packet receive timestamp.
-The following interface complements the existing ones: receive time
-stamps can be generated and returned for arbitrary packets and much
-closer to the point where the packet is really sent. Time stamps can
-be generated in software (as before) or in hardware (if the hardware
-has such a feature).
+* SO_TIMESTAMPING
+ Generates timestamps on reception, transmission or both. Supports
+ multiple timestamp sources, including hardware. Supports generating
+ timestamps for stream sockets.
-SO_TIMESTAMPING:
-Instructs the socket layer which kind of information should be collected
-and/or reported. The parameter is an integer with some of the following
-bits set. Setting other bits is an error and doesn't change the current
-state.
+1.1 SO_TIMESTAMP:
-Four of the bits are requests to the stack to try to generate
-timestamps. Any combination of them is valid.
+This socket option enables timestamping of datagrams on the reception
+path. Because the destination socket, if any, is not known early in
+the network stack, the feature has to be enabled for all packets. The
+same is true for all early receive timestamp options.
-SOF_TIMESTAMPING_TX_HARDWARE: try to obtain send time stamps in hardware
-SOF_TIMESTAMPING_TX_SOFTWARE: try to obtain send time stamps in software
-SOF_TIMESTAMPING_RX_HARDWARE: try to obtain receive time stamps in hardware
-SOF_TIMESTAMPING_RX_SOFTWARE: try to obtain receive time stamps in software
+For interface details, see `man 7 socket`.
+
+
+1.2 SO_TIMESTAMPNS:
+
+This option is identical to SO_TIMESTAMP except for the returned data type.
+Its struct timespec allows for higher resolution (ns) timestamps than the
+timeval of SO_TIMESTAMP (ms).
+
+
+1.3 SO_TIMESTAMPING:
+
+Supports multiple types of timestamp requests. As a result, this
+socket option takes a bitmap of flags, not a boolean. In
+
+ err = setsockopt(fd, SOL_SOCKET, SO_TIMESTAMPING, (void *) val, &val);
+
+val is an integer with any of the following bits set. Setting other
+bit returns EINVAL and does not change the current state.
-The other three bits control which timestamps will be reported in a
-generated control message. If none of these bits are set or if none of
-the set bits correspond to data that is available, then the control
-message will not be generated:
-SOF_TIMESTAMPING_SOFTWARE: report systime if available
-SOF_TIMESTAMPING_SYS_HARDWARE: report hwtimetrans if available (deprecated)
-SOF_TIMESTAMPING_RAW_HARDWARE: report hwtimeraw if available
+1.3.1 Timestamp Generation
-It is worth noting that timestamps may be collected for reasons other
-than being requested by a particular socket with
-SOF_TIMESTAMPING_[TR]X_(HARD|SOFT)WARE. For example, most drivers that
-can generate hardware receive timestamps ignore
-SOF_TIMESTAMPING_RX_HARDWARE. It is still a good idea to set that flag
-in case future drivers pay attention.
+Some bits are requests to the stack to try to generate timestamps. Any
+combination of them is valid. Changes to these bits apply to newly
+created packets, not to packets already in the stack. As a result, it
+is possible to selectively request timestamps for a subset of packets
+(e.g., for sampling) by embedding an send() call within two setsockopt
+calls, one to enable timestamp generation and one to disable it.
+Timestamps may also be generated for reasons other than being
+requested by a particular socket, such as when receive timestamping is
+enabled system wide, as explained earlier.
-If timestamps are reported, they will appear in a control message with
-cmsg_level==SOL_SOCKET, cmsg_type==SO_TIMESTAMPING, and a payload like
-this:
+SOF_TIMESTAMPING_RX_HARDWARE:
+ Request rx timestamps generated by the network adapter.
+
+SOF_TIMESTAMPING_RX_SOFTWARE:
+ Request rx timestamps when data enters the kernel. These timestamps
+ are generated just after a device driver hands a packet to the
+ kernel receive stack.
+
+SOF_TIMESTAMPING_TX_HARDWARE:
+ Request tx timestamps generated by the network adapter.
+
+SOF_TIMESTAMPING_TX_SOFTWARE:
+ Request tx timestamps when data leaves the kernel. These timestamps
+ are generated in the device driver as close as possible, but always
+ prior to, passing the packet to the network interface. Hence, they
+ require driver support and may not be available for all devices.
+
+SOF_TIMESTAMPING_TX_SCHED:
+ Request tx timestamps prior to entering the packet scheduler. Kernel
+ transmit latency is, if long, often dominated by queuing delay. The
+ difference between this timestamp and one taken at
+ SOF_TIMESTAMPING_TX_SOFTWARE will expose this latency independent
+ of protocol processing. The latency incurred in protocol
+ processing, if any, can be computed by subtracting a userspace
+ timestamp taken immediately before send() from this timestamp. On
+ machines with virtual devices where a transmitted packet travels
+ through multiple devices and, hence, multiple packet schedulers,
+ a timestamp is generated at each layer. This allows for fine
+ grained measurement of queuing delay.
+
+SOF_TIMESTAMPING_TX_ACK:
+ Request tx timestamps when all data in the send buffer has been
+ acknowledged. This only makes sense for reliable protocols. It is
+ currently only implemented for TCP. For that protocol, it may
+ over-report measurement, because the timestamp is generated when all
+ data up to and including the buffer at send() was acknowledged: the
+ cumulative acknowledgment. The mechanism ignores SACK and FACK.
+
+
+1.3.2 Timestamp Reporting
+
+The other three bits control which timestamps will be reported in a
+generated control message. Changes to the bits take immediate
+effect at the timestamp reporting locations in the stack. Timestamps
+are only reported for packets that also have the relevant timestamp
+generation request set.
+
+SOF_TIMESTAMPING_SOFTWARE:
+ Report any software timestamps when available.
+
+SOF_TIMESTAMPING_SYS_HARDWARE:
+ This option is deprecated and ignored.
+
+SOF_TIMESTAMPING_RAW_HARDWARE:
+ Report hardware timestamps as generated by
+ SOF_TIMESTAMPING_TX_HARDWARE when available.
+
+
+1.3.3 Timestamp Options
+
+The interface supports one option
+
+SOF_TIMESTAMPING_OPT_ID:
+
+ Generate a unique identifier along with each packet. A process can
+ have multiple concurrent timestamping requests outstanding. Packets
+ can be reordered in the transmit path, for instance in the packet
+ scheduler. In that case timestamps will be queued onto the error
+ queue out of order from the original send() calls. This option
+ embeds a counter that is incremented at send() time, to order
+ timestamps within a flow.
+
+ This option is implemented only for transmit timestamps. There, the
+ timestamp is always looped along with a struct sock_extended_err.
+ The option modifies field ee_info to pass an id that is unique
+ among all possibly concurrently outstanding timestamp requests for
+ that socket. In practice, it is a monotonically increasing u32
+ (that wraps).
+
+ In datagram sockets, the counter increments on each send call. In
+ stream sockets, it increments with every byte.
+
+
+1.4 Bytestream Timestamps
+
+The SO_TIMESTAMPING interface supports timestamping of bytes in a
+bytestream. Each request is interpreted as a request for when the
+entire contents of the buffer has passed a timestamping point. That
+is, for streams option SOF_TIMESTAMPING_TX_SOFTWARE will record
+when all bytes have reached the device driver, regardless of how
+many packets the data has been converted into.
+
+In general, bytestreams have no natural delimiters and therefore
+correlating a timestamp with data is non-trivial. A range of bytes
+may be split across segments, any segments may be merged (possibly
+coalescing sections of previously segmented buffers associated with
+independent send() calls). Segments can be reordered and the same
+byte range can coexist in multiple segments for protocols that
+implement retransmissions.
+
+It is essential that all timestamps implement the same semantics,
+regardless of these possible transformations, as otherwise they are
+incomparable. Handling "rare" corner cases differently from the
+simple case (a 1:1 mapping from buffer to skb) is insufficient
+because performance debugging often needs to focus on such outliers.
+
+In practice, timestamps can be correlated with segments of a
+bytestream consistently, if both semantics of the timestamp and the
+timing of measurement are chosen correctly. This challenge is no
+different from deciding on a strategy for IP fragmentation. There, the
+definition is that only the first fragment is timestamped. For
+bytestreams, we chose that a timestamp is generated only when all
+bytes have passed a point. SOF_TIMESTAMPING_TX_ACK as defined is easy to
+implement and reason about. An implementation that has to take into
+account SACK would be more complex due to possible transmission holes
+and out of order arrival.
+
+On the host, TCP can also break the simple 1:1 mapping from buffer to
+skbuff as a result of Nagle, cork, autocork, segmentation and GSO. The
+implementation ensures correctness in all cases by tracking the
+individual last byte passed to send(), even if it is no longer the
+last byte after an skbuff extend or merge operation. It stores the
+relevant sequence number in skb_shinfo(skb)->tskey. Because an skbuff
+has only one such field, only one timestamp can be generated.
+
+In rare cases, a timestamp request can be missed if two requests are
+collapsed onto the same skb. A process can detect this situation by
+enabling SOF_TIMESTAMPING_OPT_ID and comparing the byte offset at
+send time with the value returned for each timestamp. It can prevent
+the situation by always flushing the TCP stack in between requests,
+for instance by enabling TCP_NODELAY and disabling TCP_CORK and
+autocork.
+
+These precautions ensure that the timestamp is generated only when all
+bytes have passed a timestamp point, assuming that the network stack
+itself does not reorder the segments. The stack indeed tries to avoid
+reordering. The one exception is under administrator control: it is
+possible to construct a packet scheduler configuration that delays
+segments from the same stream differently. Such a setup would be
+unusual.
+
+
+2 Data Interfaces
+
+Timestamps are read using the ancillary data feature of recvmsg().
+See `man 3 cmsg` for details of this interface. The socket manual
+page (`man 7 socket`) describes how timestamps generated with
+SO_TIMESTAMP and SO_TIMESTAMPNS records can be retrieved.
+
+
+2.1 SCM_TIMESTAMPING records
+
+These timestamps are returned in a control message with cmsg_level
+SOL_SOCKET, cmsg_type SCM_TIMESTAMPING, and payload of type
struct scm_timestamping {
- struct timespec systime;
- struct timespec hwtimetrans;
- struct timespec hwtimeraw;
+ struct timespec ts[3];
};
-recvmsg() can be used to get this control message for regular incoming
-packets. For send time stamps the outgoing packet is looped back to
-the socket's error queue with the send time stamp(s) attached. It can
-be received with recvmsg(flags=MSG_ERRQUEUE). The call returns the
-original outgoing packet data including all headers preprended down to
-and including the link layer, the scm_timestamping control message and
-a sock_extended_err control message with ee_errno==ENOMSG and
-ee_origin==SO_EE_ORIGIN_TIMESTAMPING. A socket with such a pending
-bounced packet is ready for reading as far as select() is concerned.
-If the outgoing packet has to be fragmented, then only the first
-fragment is time stamped and returned to the sending socket.
-
-All three values correspond to the same event in time, but were
-generated in different ways. Each of these values may be empty (= all
-zero), in which case no such value was available. If the application
-is not interested in some of these values, they can be left blank to
-avoid the potential overhead of calculating them.
-
-systime is the value of the system time at that moment. This
-corresponds to the value also returned via SO_TIMESTAMP[NS]. If the
-time stamp was generated by hardware, then this field is
-empty. Otherwise it is filled in if SOF_TIMESTAMPING_SOFTWARE is
-set.
-
-hwtimeraw is the original hardware time stamp. Filled in if
-SOF_TIMESTAMPING_RAW_HARDWARE is set. No assumptions about its
-relation to system time should be made.
-
-hwtimetrans is always zero. This field is deprecated. It used to hold
-hw timestamps converted to system time. Instead, expose the hardware
-clock device on the NIC directly as a HW PTP clock source, to allow
-time conversion in userspace and optionally synchronize system time
-with a userspace PTP stack such as linuxptp. For the PTP clock API,
-see Documentation/ptp/ptp.txt.
-
-
-SIOCSHWTSTAMP, SIOCGHWTSTAMP:
+The structure can return up to three timestamps. This is a legacy
+feature. Only one field is non-zero at any time. Most timestamps
+are passed in ts[0]. Hardware timestamps are passed in ts[2].
+
+ts[1] used to hold hardware timestamps converted to system time.
+Instead, expose the hardware clock device on the NIC directly as
+a HW PTP clock source, to allow time conversion in userspace and
+optionally synchronize system time with a userspace PTP stack such
+as linuxptp. For the PTP clock API, see Documentation/ptp/ptp.txt.
+
+2.1.1 Transmit timestamps with MSG_ERRQUEUE
+
+For transmit timestamps the outgoing packet is looped back to the
+socket's error queue with the send timestamp(s) attached. A process
+receives the timestamps by calling recvmsg() with flag MSG_ERRQUEUE
+set and with a msg_control buffer sufficiently large to receive the
+relevant metadata structures. The recvmsg call returns the original
+outgoing data packet with two ancillary messages attached.
+
+A message of cm_level SOL_IP(V6) and cm_type IP(V6)_RECVERR
+embeds a struct sock_extended_err. This defines the error type. For
+timestamps, the ee_errno field is ENOMSG. The other ancillary message
+will have cm_level SOL_SOCKET and cm_type SCM_TIMESTAMPING. This
+embeds the struct scm_timestamping.
+
+
+2.1.1.2 Timestamp types
+
+The semantics of the three struct timespec are defined by field
+ee_info in the extended error structure. It contains a value of
+type SCM_TSTAMP_* to define the actual timestamp passed in
+scm_timestamping.
+
+The SCM_TSTAMP_* types are 1:1 matches to the SOF_TIMESTAMPING_*
+control fields discussed previously, with one exception. For legacy
+reasons, SCM_TSTAMP_SND is equal to zero and can be set for both
+SOF_TIMESTAMPING_TX_HARDWARE and SOF_TIMESTAMPING_TX_SOFTWARE. It
+is the first if ts[2] is non-zero, the second otherwise, in which
+case the timestamp is stored in ts[0].
+
+
+2.1.1.3 Fragmentation
+
+Fragmentation of outgoing datagrams is rare, but is possible, e.g., by
+explicitly disabling PMTU discovery. If an outgoing packet is fragmented,
+then only the first fragment is timestamped and returned to the sending
+socket.
+
+
+2.1.1.4 Packet Payload
+
+The calling application is often not interested in receiving the whole
+packet payload that it passed to the stack originally: the socket
+error queue mechanism is just a method to piggyback the timestamp on.
+In this case, the application can choose to read datagrams with a
+smaller buffer, possibly even of length 0. The payload is truncated
+accordingly. Until the process calls recvmsg() on the error queue,
+however, the full packet is queued, taking up budget from SO_RCVBUF.
+
+
+2.1.1.5 Blocking Read
+
+Reading from the error queue is always a non-blocking operation. To
+block waiting on a timestamp, use poll or select. poll() will return
+POLLERR in pollfd.revents if any data is ready on the error queue.
+There is no need to pass this flag in pollfd.events. This flag is
+ignored on request. See also `man 2 poll`.
+
+
+2.1.2 Receive timestamps
+
+On reception, there is no reason to read from the socket error queue.
+The SCM_TIMESTAMPING ancillary data is sent along with the packet data
+on a normal recvmsg(). Since this is not a socket error, it is not
+accompanied by a message SOL_IP(V6)/IP(V6)_RECVERROR. In this case,
+the meaning of the three fields in struct scm_timestamping is
+implicitly defined. ts[0] holds a software timestamp if set, ts[1]
+is again deprecated and ts[2] holds a hardware timestamp if set.
+
+
+3. Hardware Timestamping configuration: SIOCSHWTSTAMP and SIOCGHWTSTAMP
Hardware time stamping must also be initialized for each device driver
that is expected to do hardware time stamping. The parameter is defined in
@@ -167,8 +372,7 @@ enum {
*/
};
-
-DEVICE IMPLEMENTATION
+3.1 Hardware Timestamping Implementation: Device Drivers
A driver which supports hardware time stamping must support the
SIOCSHWTSTAMP ioctl and update the supplied struct hwtstamp_config with
diff --git a/Documentation/networking/timestamping/Makefile b/Documentation/networking/timestamping/Makefile
index d934afc..95e239c 100644
--- a/Documentation/networking/timestamping/Makefile
+++ b/Documentation/networking/timestamping/Makefile
@@ -1,14 +1,20 @@
+# To compile, from the source root
+#
+# make headers_install
+# make M=documentation
+
# kbuild trick to avoid linker error. Can be omitted if a module is built.
obj- := dummy.o
# List of programs to build
-hostprogs-y := timestamping hwtstamp_config
+hostprogs-y := timestamping txtimestamp hwtstamp_config
# Tell kbuild to always build the programs
always := $(hostprogs-y)
HOSTCFLAGS_timestamping.o += -I$(objtree)/usr/include
+HOSTCFLAGS_txtimestamp.o += -I$(objtree)/usr/include
HOSTCFLAGS_hwtstamp_config.o += -I$(objtree)/usr/include
clean:
- rm -f timestamping hwtstamp_config
+ rm -f timestamping txtimestamp hwtstamp_config
diff --git a/Documentation/networking/timestamping/txtimestamp.c b/Documentation/networking/timestamping/txtimestamp.c
new file mode 100644
index 0000000..e5b0b98
--- /dev/null
+++ b/Documentation/networking/timestamping/txtimestamp.c
@@ -0,0 +1,470 @@
+/*
+ * Copyright 2014 Google Inc.
+ * Author: willemb@google.com (Willem de Bruijn)
+ *
+ * Test software tx timestamping, including
+ *
+ * - SCHED, SND and ACK timestamps
+ * - RAW, UDP and TCP
+ * - IPv4 and IPv6
+ * - various packet sizes (to test GSO and TSO)
+ *
+ * Consult the command line arguments for help on running
+ * the various testcases.
+ *
+ * This test requires a dummy TCP server.
+ * A simple `nc6 [-u] -l -p $DESTPORT` will do
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. * See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <arpa/inet.h>
+#include <asm/types.h>
+#include <error.h>
+#include <errno.h>
+#include <linux/errqueue.h>
+#include <linux/if_ether.h>
+#include <linux/net_tstamp.h>
+#include <netdb.h>
+#include <net/if.h>
+#include <netinet/in.h>
+#include <netinet/ip.h>
+#include <netinet/udp.h>
+#include <netinet/tcp.h>
+#include <netpacket/packet.h>
+#include <poll.h>
+#include <stdarg.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <sys/select.h>
+#include <sys/socket.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <time.h>
+#include <unistd.h>
+
+/* command line parameters */
+static int cfg_proto = SOCK_STREAM;
+static int cfg_ipproto = IPPROTO_TCP;
+static int cfg_num_pkts = 4;
+static int do_ipv4 = 1;
+static int do_ipv6 = 1;
+static int cfg_payload_len = 10;
+static uint16_t dest_port = 9000;
+
+static struct sockaddr_in daddr;
+static struct sockaddr_in6 daddr6;
+static struct timespec ts_prev;
+
+static void __print_timestamp(const char *name, struct timespec *cur,
+ uint32_t key, int payload_len)
+{
+ if (!(cur->tv_sec | cur->tv_nsec))
+ return;
+
+ fprintf(stderr, " %s: %lu s %lu us (seq=%u, len=%u)",
+ name, cur->tv_sec, cur->tv_nsec / 1000,
+ key, payload_len);
+
+ if ((ts_prev.tv_sec | ts_prev.tv_nsec)) {
+ int64_t cur_ms, prev_ms;
+
+ cur_ms = (long) cur->tv_sec * 1000 * 1000;
+ cur_ms += cur->tv_nsec / 1000;
+
+ prev_ms = (long) ts_prev.tv_sec * 1000 * 1000;
+ prev_ms += ts_prev.tv_nsec / 1000;
+
+ fprintf(stderr, " (%+ld us)", cur_ms - prev_ms);
+ }
+
+ ts_prev = *cur;
+ fprintf(stderr, "\n");
+}
+
+static void print_timestamp_usr(void)
+{
+ struct timespec ts;
+ struct timeval tv; /* avoid dependency on -lrt */
+
+ gettimeofday(&tv, NULL);
+ ts.tv_sec = tv.tv_sec;
+ ts.tv_nsec = tv.tv_usec * 1000;
+
+ __print_timestamp(" USR", &ts, 0, 0);
+}
+
+static void print_timestamp(struct scm_timestamping *tss, int tstype,
+ int tskey, int payload_len)
+{
+ const char *tsname;
+
+ switch (tstype) {
+ case SCM_TSTAMP_SCHED:
+ tsname = " ENQ";
+ break;
+ case SCM_TSTAMP_SND:
+ tsname = " SND";
+ break;
+ case SCM_TSTAMP_ACK:
+ tsname = " ACK";
+ break;
+ default:
+ error(1, 0, "unknown timestamp type: %u",
+ tstype);
+ }
+ __print_timestamp(tsname, &tss->ts[0], tskey, payload_len);
+}
+
+static void __poll(int fd)
+{
+ struct pollfd pollfd;
+ int ret;
+
+ memset(&pollfd, 0, sizeof(pollfd));
+ pollfd.fd = fd;
+ ret = poll(&pollfd, 1, 100);
+ if (ret != 1)
+ error(1, errno, "poll");
+}
+
+static void __recv_errmsg_cmsg(struct msghdr *msg, int payload_len)
+{
+ struct sock_extended_err *serr = NULL;
+ struct scm_timestamping *tss = NULL;
+ struct cmsghdr *cm;
+
+ for (cm = CMSG_FIRSTHDR(msg);
+ cm && cm->cmsg_len;
+ cm = CMSG_NXTHDR(msg, cm)) {
+ if (cm->cmsg_level == SOL_SOCKET &&
+ cm->cmsg_type == SCM_TIMESTAMPING) {
+ tss = (void *) CMSG_DATA(cm);
+ } else if ((cm->cmsg_level == SOL_IP &&
+ cm->cmsg_type == IP_RECVERR) ||
+ (cm->cmsg_level == SOL_IPV6 &&
+ cm->cmsg_type == IPV6_RECVERR)) {
+
+ serr = (void *) CMSG_DATA(cm);
+ if (serr->ee_errno != ENOMSG ||
+ serr->ee_origin != SO_EE_ORIGIN_TIMESTAMPING) {
+ fprintf(stderr, "unknown ip error %d %d\n",
+ serr->ee_errno,
+ serr->ee_origin);
+ serr = NULL;
+ }
+ } else
+ fprintf(stderr, "unknown cmsg %d,%d\n",
+ cm->cmsg_level, cm->cmsg_type);
+ }
+
+ if (serr && tss)
+ print_timestamp(tss, serr->ee_info, serr->ee_data, payload_len);
+}
+
+static int recv_errmsg(int fd)
+{
+ static char ctrl[1024 /* overprovision*/];
+ static struct msghdr msg;
+ struct iovec entry;
+ static char *data;
+ int ret = 0;
+
+ data = malloc(cfg_payload_len);
+ if (!data)
+ error(1, 0, "malloc");
+
+ memset(&msg, 0, sizeof(msg));
+ memset(&entry, 0, sizeof(entry));
+ memset(ctrl, 0, sizeof(ctrl));
+ memset(data, 0, sizeof(data));
+
+ entry.iov_base = data;
+ entry.iov_len = cfg_payload_len;
+ msg.msg_iov = &entry;
+ msg.msg_iovlen = 1;
+ msg.msg_name = NULL;
+ msg.msg_namelen = 0;
+ msg.msg_control = ctrl;
+ msg.msg_controllen = sizeof(ctrl);
+
+ ret = recvmsg(fd, &msg, MSG_ERRQUEUE);
+ if (ret == -1 && errno != EAGAIN)
+ error(1, errno, "recvmsg");
+
+ __recv_errmsg_cmsg(&msg, ret);
+
+ free(data);
+ return ret == -1;
+}
+
+static void do_test(int family, unsigned int opt)
+{
+ char *buf;
+ int fd, i, val, total_len;
+
+ if (family == IPPROTO_IPV6 && cfg_proto != SOCK_STREAM) {
+ /* due to lack of checksum generation code */
+ fprintf(stderr, "test: skipping datagram over IPv6\n");
+ return;
+ }
+
+ total_len = cfg_payload_len;
+ if (cfg_proto == SOCK_RAW) {
+ total_len += sizeof(struct udphdr);
+ if (cfg_ipproto == IPPROTO_RAW)
+ total_len += sizeof(struct iphdr);
+ }
+
+ buf = malloc(total_len);
+ if (!buf)
+ error(1, 0, "malloc");
+
+ fd = socket(family, cfg_proto, cfg_ipproto);
+ if (fd < 0)
+ error(1, errno, "socket");
+
+ if (cfg_proto == SOCK_STREAM) {
+ val = 1;
+ if (setsockopt(fd, IPPROTO_TCP, TCP_NODELAY,
+ (char*) &val, sizeof(val)))
+ error(1, 0, "setsockopt no nagle");
+
+ if (family == PF_INET) {
+ if (connect(fd, (void *) &daddr, sizeof(daddr)))
+ error(1, errno, "connect ipv4");
+ } else {
+ if (connect(fd, (void *) &daddr6, sizeof(daddr6)))
+ error(1, errno, "connect ipv6");
+ }
+ }
+
+ opt |= SOF_TIMESTAMPING_SOFTWARE |
+ SOF_TIMESTAMPING_OPT_ID;
+ if (setsockopt(fd, SOL_SOCKET, SO_TIMESTAMPING,
+ (char *) &opt, sizeof(opt)))
+ error(1, 0, "setsockopt timestamping");
+
+ for (i = 0; i < cfg_num_pkts; i++) {
+ memset(&ts_prev, 0, sizeof(ts_prev));
+ memset(buf, 'a' + i, total_len);
+ buf[total_len - 2] = '\n';
+ buf[total_len - 1] = '\0';
+
+ if (cfg_proto == SOCK_RAW) {
+ struct udphdr *udph;
+ int off = 0;
+
+ if (cfg_ipproto == IPPROTO_RAW) {
+ struct iphdr *iph = (void *) buf;
+
+ memset(iph, 0, sizeof(*iph));
+ iph->ihl = 5;
+ iph->version = 4;
+ iph->ttl = 2;
+ iph->daddr = daddr.sin_addr.s_addr;
+ iph->protocol = IPPROTO_UDP;
+ /* kernel writes saddr, csum, len */
+
+ off = sizeof(*iph);
+ }
+
+ udph = (void *) buf + off;
+ udph->source = ntohs(9000); /* random spoof */
+ udph->dest = ntohs(dest_port);
+ udph->len = ntohs(sizeof(*udph) + cfg_payload_len);
+ udph->check = 0; /* not allowed for IPv6 */
+ }
+
+ print_timestamp_usr();
+ if (cfg_proto != SOCK_STREAM) {
+ if (family == PF_INET)
+ val = sendto(fd, buf, total_len, 0, (void *) &daddr, sizeof(daddr));
+ else
+ val = sendto(fd, buf, total_len, 0, (void *) &daddr6, sizeof(daddr6));
+ } else {
+ val = send(fd, buf, cfg_payload_len, 0);
+ }
+ if (val != total_len)
+ error(1, errno, "send");
+
+ /* wait for all errors to be queued, else ACKs arrive OOO */
+ usleep(50 * 1000);
+
+ __poll(fd);
+
+ while (!recv_errmsg(fd)) {}
+ }
+
+ if (close(fd))
+ error(1, errno, "close");
+
+ free(buf);
+ usleep(400 * 1000);
+}
+
+static void __attribute__((noreturn)) usage(const char *filepath)
+{
+ fprintf(stderr, "\nUsage: %s [options] hostname\n"
+ "\nwhere options are:\n"
+ " -4: only IPv4\n"
+ " -6: only IPv6\n"
+ " -h: show this message\n"
+ " -l N: send N bytes at a time\n"
+ " -r: use raw\n"
+ " -R: use raw (IP_HDRINCL)\n"
+ " -p N: connect to port N\n"
+ " -u: use udp\n",
+ filepath);
+ exit(1);
+}
+
+static void parse_opt(int argc, char **argv)
+{
+ int proto_count = 0;
+ char c;
+
+ while ((c = getopt(argc, argv, "46hl:p:rRu")) != -1) {
+ switch (c) {
+ case '4':
+ do_ipv6 = 0;
+ break;
+ case '6':
+ do_ipv4 = 0;
+ break;
+ case 'r':
+ proto_count++;
+ cfg_proto = SOCK_RAW;
+ cfg_ipproto = IPPROTO_UDP;
+ break;
+ case 'R':
+ proto_count++;
+ cfg_proto = SOCK_RAW;
+ cfg_ipproto = IPPROTO_RAW;
+ break;
+ case 'u':
+ proto_count++;
+ cfg_proto = SOCK_DGRAM;
+ cfg_ipproto = IPPROTO_UDP;
+ break;
+ case 'l':
+ cfg_payload_len = strtoul(optarg, NULL, 10);
+ break;
+ case 'p':
+ dest_port = strtoul(optarg, NULL, 10);
+ break;
+ case 'h':
+ default:
+ usage(argv[0]);
+ }
+ }
+
+ if (!cfg_payload_len)
+ error(1, 0, "payload may not be nonzero");
+ if (cfg_proto != SOCK_STREAM && cfg_payload_len > 1472)
+ error(1, 0, "udp packet might exceed expected MTU");
+ if (!do_ipv4 && !do_ipv6)
+ error(1, 0, "pass -4 or -6, not both");
+ if (proto_count > 1)
+ error(1, 0, "pass -r, -R or -u, not multiple");
+
+ if (optind != argc - 1)
+ error(1, 0, "missing required hostname argument");
+}
+
+static void resolve_hostname(const char *hostname)
+{
+ struct addrinfo *addrs, *cur;
+ int have_ipv4 = 0, have_ipv6 = 0;
+
+ if (getaddrinfo(hostname, NULL, NULL, &addrs))
+ error(1, errno, "getaddrinfo");
+
+ cur = addrs;
+ while (cur && !have_ipv4 && !have_ipv6) {
+ if (!have_ipv4 && cur->ai_family == AF_INET) {
+ memcpy(&daddr, cur->ai_addr, sizeof(daddr));
+ daddr.sin_port = htons(dest_port);
+ have_ipv4 = 1;
+ }
+ else if (!have_ipv6 && cur->ai_family == AF_INET6) {
+ memcpy(&daddr6, cur->ai_addr, sizeof(daddr6));
+ daddr6.sin6_port = htons(dest_port);
+ have_ipv6 = 1;
+ }
+ cur = cur->ai_next;
+ }
+ if (addrs)
+ freeaddrinfo(addrs);
+
+ do_ipv4 &= have_ipv4;
+ do_ipv6 &= have_ipv6;
+}
+
+static void do_main(int family)
+{
+ fprintf(stderr, "family: %s\n",
+ family == PF_INET ? "INET" : "INET6");
+
+ fprintf(stderr, "test SND\n");
+ do_test(family, SOF_TIMESTAMPING_TX_SOFTWARE);
+
+ fprintf(stderr, "test ENQ\n");
+ do_test(family, SOF_TIMESTAMPING_TX_SCHED);
+
+ fprintf(stderr, "test ENQ + SND\n");
+ do_test(family, SOF_TIMESTAMPING_TX_SCHED |
+ SOF_TIMESTAMPING_TX_SOFTWARE);
+
+ if (cfg_proto == SOCK_STREAM) {
+ fprintf(stderr, "\ntest ACK\n");
+ do_test(family, SOF_TIMESTAMPING_TX_ACK);
+
+ fprintf(stderr, "\ntest SND + ACK\n");
+ do_test(family, SOF_TIMESTAMPING_TX_SOFTWARE |
+ SOF_TIMESTAMPING_TX_ACK);
+
+ fprintf(stderr, "\ntest ENQ + SND + ACK\n");
+ do_test(family, SOF_TIMESTAMPING_TX_SCHED |
+ SOF_TIMESTAMPING_TX_SOFTWARE |
+ SOF_TIMESTAMPING_TX_ACK);
+ }
+}
+
+const char *sock_names[] = { NULL, "TCP", "UDP", "RAW" };
+
+int main(int argc, char **argv)
+{
+ if (argc == 1)
+ usage(argv[0]);
+
+ parse_opt(argc, argv);
+ resolve_hostname(argv[argc - 1]);
+
+ fprintf(stderr, "protocol: %s\n", sock_names[cfg_proto]);
+ fprintf(stderr, "payload: %u\n", cfg_payload_len);
+ fprintf(stderr, "server port: %u\n", dest_port);
+ fprintf(stderr, "\n");
+
+ if (do_ipv4)
+ do_main(PF_INET);
+ if (do_ipv6)
+ do_main(PF_INET6);
+
+ return 0;
+}
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next 6/6] l2tp: Enable checksum unnecessary conversions for l2tp/UDP sockets
From: Tom Herbert @ 2014-08-31 22:12 UTC (permalink / raw)
To: davem, netdev
In-Reply-To: <1409523166-9215-1-git-send-email-therbert@google.com>
Signed-off-by: Tom Herbert <therbert@google.com>
---
net/l2tp/l2tp_core.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c
index 1109d3b..797c0af 100644
--- a/net/l2tp/l2tp_core.c
+++ b/net/l2tp/l2tp_core.c
@@ -1392,6 +1392,8 @@ static int l2tp_tunnel_sock_create(struct net *net,
if (err < 0)
goto out;
+ udp_set_convert_csum(sock->sk, true);
+
break;
case L2TP_ENCAPTYPE_IP:
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next 5/6] vxlan: Enable checksum unnecessary conversions for vxlan/UDP sockets
From: Tom Herbert @ 2014-08-31 22:12 UTC (permalink / raw)
To: davem, netdev
In-Reply-To: <1409523166-9215-1-git-send-email-therbert@google.com>
Signed-off-by: Tom Herbert <therbert@google.com>
---
drivers/net/vxlan.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 67527f3..53c3ec1 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -2370,6 +2370,8 @@ static struct socket *vxlan_create_sock(struct net *net, bool ipv6,
/* Disable multicast loopback */
inet_sk(sock->sk)->mc_loop = 0;
+ udp_set_convert_csum(sock->sk, true);
+
return sock;
}
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next 4/6] gre: Add support for checksum unnecessary conversions
From: Tom Herbert @ 2014-08-31 22:12 UTC (permalink / raw)
To: davem, netdev
In-Reply-To: <1409523166-9215-1-git-send-email-therbert@google.com>
Call skb_checksum_try_convert and skb_gro_checksum_try_convert
after checksum is found present and validated in the GRE header
for normal and GRO paths respectively.
In GRO path, call skb_gro_checksum_try_convert
Signed-off-by: Tom Herbert <therbert@google.com>
---
net/ipv4/gre_demux.c | 4 ++++
net/ipv4/gre_offload.c | 8 ++++++--
2 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/gre_demux.c b/net/ipv4/gre_demux.c
index 0485bf7..7e0756d 100644
--- a/net/ipv4/gre_demux.c
+++ b/net/ipv4/gre_demux.c
@@ -125,6 +125,10 @@ static int parse_gre_header(struct sk_buff *skb, struct tnl_ptk_info *tpi,
*csum_err = true;
return -EINVAL;
}
+
+ skb_checksum_try_convert(skb, IPPROTO_GRE, 0,
+ null_compute_pseudo);
+
options++;
}
diff --git a/net/ipv4/gre_offload.c b/net/ipv4/gre_offload.c
index a4d7965..d3fe2ac 100644
--- a/net/ipv4/gre_offload.c
+++ b/net/ipv4/gre_offload.c
@@ -172,10 +172,14 @@ static struct sk_buff **gre_gro_receive(struct sk_buff **head,
}
/* Don't bother verifying checksum if we're going to flush anyway. */
- if ((greh->flags & GRE_CSUM) && !NAPI_GRO_CB(skb)->flush &&
- skb_gro_checksum_simple_validate(skb))
+ if ((greh->flags & GRE_CSUM) && !NAPI_GRO_CB(skb)->flush) {
+ if (skb_gro_checksum_simple_validate(skb))
goto out_unlock;
+ skb_gro_checksum_try_convert(skb, IPPROTO_GRE, 0,
+ null_compute_pseudo);
+ }
+
flush = 0;
for (p = *head; p; p = p->next) {
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next 3/6] udp: Add support for doing checksum unnecessary conversion
From: Tom Herbert @ 2014-08-31 22:12 UTC (permalink / raw)
To: davem, netdev
In-Reply-To: <1409523166-9215-1-git-send-email-therbert@google.com>
Add support for doing CHECKSUM_UNNECESSARY to CHECKSUM_COMPLETE
conversion in UDP tunneling path.
In the normal UDP path, we call skb_checksum_try_convert after locating
the UDP socket. The check is that checksum conversion is enabled for
the socket (new flag in UDP socket) and that checksum field is
non-zero.
In the UDP GRO path, we call skb_gro_checksum_try_convert after
checksum is validated and checksum field is non-zero. Since this is
already in GRO we assume that checksum conversion is always wanted.
Signed-off-by: Tom Herbert <therbert@google.com>
---
include/linux/udp.h | 16 +++++++++++++++-
net/ipv4/udp.c | 4 ++++
net/ipv4/udp_offload.c | 25 +++++++++++++++++--------
net/ipv6/udp.c | 4 ++++
net/ipv6/udp_offload.c | 24 +++++++++++++++++-------
5 files changed, 57 insertions(+), 16 deletions(-)
diff --git a/include/linux/udp.h b/include/linux/udp.h
index 247cfdc..ee32775 100644
--- a/include/linux/udp.h
+++ b/include/linux/udp.h
@@ -49,7 +49,11 @@ struct udp_sock {
unsigned int corkflag; /* Cork is required */
__u8 encap_type; /* Is this an Encapsulation socket? */
unsigned char no_check6_tx:1,/* Send zero UDP6 checksums on TX? */
- no_check6_rx:1;/* Allow zero UDP6 checksums on RX? */
+ no_check6_rx:1,/* Allow zero UDP6 checksums on RX? */
+ convert_csum:1;/* On receive, convert checksum
+ * unnecessary to checksum complete
+ * if possible.
+ */
/*
* Following member retains the information to create a UDP header
* when the socket is uncorked.
@@ -98,6 +102,16 @@ static inline bool udp_get_no_check6_rx(struct sock *sk)
return udp_sk(sk)->no_check6_rx;
}
+static inline void udp_set_convert_csum(struct sock *sk, bool val)
+{
+ udp_sk(sk)->convert_csum = val;
+}
+
+static inline bool udp_get_convert_csum(struct sock *sk)
+{
+ return udp_sk(sk)->convert_csum;
+}
+
#define udp_portaddr_for_each_entry(__sk, node, list) \
hlist_nulls_for_each_entry(__sk, node, list, __sk_common.skc_portaddr_node)
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 3549c21..0da3849 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -1788,6 +1788,10 @@ int __udp4_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
if (sk != NULL) {
int ret;
+ if (udp_sk(sk)->convert_csum && uh->check && !IS_UDPLITE(sk))
+ skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
+ inet_compute_pseudo);
+
ret = udp_queue_rcv_skb(sk, skb);
sock_put(sk);
diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index a6adff9..84e0e05 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -290,16 +290,25 @@ static struct sk_buff **udp4_gro_receive(struct sk_buff **head,
{
struct udphdr *uh = udp_gro_udphdr(skb);
- /* Don't bother verifying checksum if we're going to flush anyway. */
- if (unlikely(!uh) ||
- (!NAPI_GRO_CB(skb)->flush &&
- skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
- inet_gro_compute_pseudo))) {
- NAPI_GRO_CB(skb)->flush = 1;
- return NULL;
- }
+ if (unlikely(!uh))
+ goto flush;
+ /* Don't bother verifying checksum if we're going to flush anyway. */
+ if (!NAPI_GRO_CB(skb)->flush)
+ goto skip;
+
+ if (skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
+ inet_gro_compute_pseudo))
+ goto flush;
+ else if (uh->check)
+ skb_gro_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
+ inet_gro_compute_pseudo);
+skip:
return udp_gro_receive(head, skb, uh);
+
+flush:
+ NAPI_GRO_CB(skb)->flush = 1;
+ return NULL;
}
int udp_gro_complete(struct sk_buff *skb, int nhoff)
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 12fcce8f..f6ba535 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -891,6 +891,10 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
goto csum_error;
}
+ if (udp_sk(sk)->convert_csum && uh->check && !IS_UDPLITE(sk))
+ skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
+ ip6_compute_pseudo);
+
ret = udpv6_queue_rcv_skb(sk, skb);
sock_put(sk);
diff --git a/net/ipv6/udp_offload.c b/net/ipv6/udp_offload.c
index b13e377..89cb9a9 100644
--- a/net/ipv6/udp_offload.c
+++ b/net/ipv6/udp_offload.c
@@ -134,16 +134,26 @@ static struct sk_buff **udp6_gro_receive(struct sk_buff **head,
{
struct udphdr *uh = udp_gro_udphdr(skb);
+ if (unlikely(!uh))
+ goto flush;
+
/* Don't bother verifying checksum if we're going to flush anyway. */
- if (unlikely(!uh) ||
- (!NAPI_GRO_CB(skb)->flush &&
- skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
- ip6_gro_compute_pseudo))) {
- NAPI_GRO_CB(skb)->flush = 1;
- return NULL;
- }
+ if (!NAPI_GRO_CB(skb)->flush)
+ goto skip;
+ if (skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
+ ip6_gro_compute_pseudo))
+ goto flush;
+ else if (uh->check)
+ skb_gro_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
+ ip6_gro_compute_pseudo);
+
+skip:
return udp_gro_receive(head, skb, uh);
+
+flush:
+ NAPI_GRO_CB(skb)->flush = 1;
+ return NULL;
}
int udp6_gro_complete(struct sk_buff *skb, int nhoff)
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next 2/6] net: Infrastructure for checksum unnecessary conversions
From: Tom Herbert @ 2014-08-31 22:12 UTC (permalink / raw)
To: davem, netdev
In-Reply-To: <1409523166-9215-1-git-send-email-therbert@google.com>
For normal path, added skb_checksum_try_convert which is called
to attempt to convert CHECKSUM_UNNECESSARY to CHECKSUM_COMPLETE. The
primary condition to allow this is that ip_summed is CHECKSUM_NONE
and csum_valid is true, which will be the state after consuming
a CHECKSUM_UNNECESSARY.
For GRO path, added skb_gro_checksum_try_convert which is the GRO
analogue of skb_checksum_try_convert. The primary condition to allow
this is that NAPI_GRO_CB(skb)->csum_cnt == 0 and
NAPI_GRO_CB(skb)->csum_valid is set. This implies that we have consumed
all available CHECKSUM_UNNECESSARY checksums in the GRO path.
Signed-off-by: Tom Herbert <therbert@google.com>
---
include/linux/netdevice.h | 20 ++++++++++++++++++++
include/linux/skbuff.h | 20 ++++++++++++++++++++
2 files changed, 40 insertions(+)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 024d8c1..9d4ab3e 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -2233,6 +2233,26 @@ static inline void skb_gro_incr_csum_unnecessary(struct sk_buff *skb)
#define skb_gro_checksum_simple_validate(skb) \
__skb_gro_checksum_validate(skb, 0, false, 0, null_compute_pseudo)
+static inline bool __skb_gro_checksum_convert_check(struct sk_buff *skb)
+{
+ return (NAPI_GRO_CB(skb)->csum_cnt == 0 &&
+ !NAPI_GRO_CB(skb)->csum_valid);
+}
+
+static inline void __skb_gro_checksum_convert(struct sk_buff *skb,
+ __sum16 check, __wsum pseudo)
+{
+ NAPI_GRO_CB(skb)->csum = ~pseudo;
+ NAPI_GRO_CB(skb)->csum_valid = 1;
+}
+
+#define skb_gro_checksum_try_convert(skb, proto, check, compute_pseudo) \
+do { \
+ if (__skb_gro_checksum_convert_check(skb)) \
+ __skb_gro_checksum_convert(skb, check, \
+ compute_pseudo(skb, proto)); \
+} while (0)
+
static inline int dev_hard_header(struct sk_buff *skb, struct net_device *dev,
unsigned short type,
const void *daddr, const void *saddr,
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 23710a2..02529fc 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -2942,6 +2942,26 @@ static inline __wsum null_compute_pseudo(struct sk_buff *skb, int proto)
#define skb_checksum_simple_validate(skb) \
__skb_checksum_validate(skb, 0, true, false, 0, null_compute_pseudo)
+static inline bool __skb_checksum_convert_check(struct sk_buff *skb)
+{
+ return (skb->ip_summed == CHECKSUM_NONE &&
+ skb->csum_valid && !skb->csum_bad);
+}
+
+static inline void __skb_checksum_convert(struct sk_buff *skb,
+ __sum16 check, __wsum pseudo)
+{
+ skb->csum = ~pseudo;
+ skb->ip_summed = CHECKSUM_COMPLETE;
+}
+
+#define skb_checksum_try_convert(skb, proto, check, compute_pseudo) \
+do { \
+ if (__skb_checksum_convert_check(skb)) \
+ __skb_checksum_convert(skb, check, \
+ compute_pseudo(skb, proto)); \
+} while (0)
+
#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
void nf_conntrack_destroy(struct nf_conntrack *nfct);
static inline void nf_conntrack_put(struct nf_conntrack *nfct)
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
* [PATCH net-next 1/6] net: Support for csum_bad in skbuff
From: Tom Herbert @ 2014-08-31 22:12 UTC (permalink / raw)
To: davem, netdev
In-Reply-To: <1409523166-9215-1-git-send-email-therbert@google.com>
This flag indicates that an invalid checksum was detected in the
packet. __skb_mark_checksum_bad helper function was added to set this.
Checksums can be marked bad from a driver or the GRO path (the latter
is implemented in this patch). csum_bad is checked in
__skb_checksum_validate_complete (i.e. calling that when ip_summed ==
CHECKSUM_NONE).
csum_bad works in conjunction with ip_summed value. In the case that
ip_summed is CHECKSUM_NONE and csum_bad is set, this implies that the
first (or next) checksum encountered in the packet is bad. When
ip_summed is CHECKSUM_UNNECESSARY, the first checksum after the last
one validated is bad. For example, if ip_summed == CHECKSUM_UNNECESSARY,
csum_level == 1, and csum_bad is set-- then the third checksum in the
packet is bad. In the normal path, the packet will be dropped when
processing the protocol layer of the bad checksum:
__skb_decr_checksum_unnecessary called twice for the good checksums
changing ip_summed to CHECKSUM_NONE so that
__skb_checksum_validate_complete is called to validate the third
checksum and that will fail since csum_bad is set.
Signed-off-by: Tom Herbert <therbert@google.com>
---
include/linux/netdevice.h | 4 +++-
include/linux/skbuff.h | 21 ++++++++++++++++++++-
net/core/dev.c | 2 +-
3 files changed, 24 insertions(+), 3 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 456eb1f..024d8c1 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -2216,7 +2216,9 @@ static inline void skb_gro_incr_csum_unnecessary(struct sk_buff *skb)
if (__skb_gro_checksum_validate_needed(skb, zero_okay, check)) \
__ret = __skb_gro_checksum_validate_complete(skb, \
compute_pseudo(skb, proto)); \
- if (!__ret) \
+ if (__ret) \
+ __skb_mark_checksum_bad(skb); \
+ else \
skb_gro_incr_csum_unnecessary(skb); \
__ret; \
})
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index c93b585..23710a2 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -617,7 +617,8 @@ struct sk_buff {
kmemcheck_bitfield_begin(flags3);
__u8 csum_level:2;
- /* 14 bit hole */
+ __u8 csum_bad:1;
+ /* 13 bit hole */
kmemcheck_bitfield_end(flags3);
__be16 inner_protocol;
@@ -2825,6 +2826,21 @@ static inline void __skb_incr_checksum_unnecessary(struct sk_buff *skb)
}
}
+static inline void __skb_mark_checksum_bad(struct sk_buff *skb)
+{
+ /* Mark current checksum as bad (typically called from GRO
+ * path). In the case that ip_summed is CHECKSUM_NONE
+ * this must be the first checksum encountered in the packet.
+ * When ip_summed is CHECKSUM_UNNECESSARY, this is the first
+ * checksum after the last one validated. For UDP, a zero
+ * checksum can not be marked as bad.
+ */
+
+ if (skb->ip_summed == CHECKSUM_NONE ||
+ skb->ip_summed == CHECKSUM_UNNECESSARY)
+ skb->csum_bad = 1;
+}
+
/* Check if we need to perform checksum complete validation.
*
* Returns true if checksum complete is needed, false otherwise
@@ -2866,6 +2882,9 @@ static inline __sum16 __skb_checksum_validate_complete(struct sk_buff *skb,
skb->csum_valid = 1;
return 0;
}
+ } else if (skb->csum_bad) {
+ /* ip_summed == CHECKSUM_NONE in this case */
+ return 1;
}
skb->csum = psum;
diff --git a/net/core/dev.c b/net/core/dev.c
index a6077ef..960df22 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3957,7 +3957,7 @@ static enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff
if (!(skb->dev->features & NETIF_F_GRO))
goto normal;
- if (skb_is_gso(skb) || skb_has_frag_list(skb))
+ if (skb_is_gso(skb) || skb_has_frag_list(skb) || skb->csum_bad)
goto normal;
gro_list_prepare(napi, skb);
--
2.1.0.rc2.206.gedb03e5
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox